Innovations in Biomedical Engineering (Lecture Notes in Networks and Systems, 409) 3030991113, 9783030991111

Citation preview

Lecture Notes in Networks and Systems 409

Marek Gzik · Zbigniew Paszenda · Ewa Piętka · Ewaryst Tkacz · Krzysztof Milewski · Jacek Jurkojć   Editors

Innovations in Biomedical Engineering

Lecture Notes in Networks and Systems Volume 409

Series Editor Janusz Kacprzyk, Systems Research Institute, Polish Academy of Sciences, Warsaw, Poland Advisory Editors Fernando Gomide, Department of Computer Engineering and Automation—DCA, School of Electrical and Computer Engineering—FEEC, University of Campinas— UNICAMP, São Paulo, Brazil Okyay Kaynak, Department of Electrical and Electronic Engineering, Bogazici University, Istanbul, Turkey Derong Liu, Department of Electrical and Computer Engineering, University of Illinois at Chicago, Chicago, USA Institute of Automation, Chinese Academy of Sciences, Beijing, China Witold Pedrycz, Department of Electrical and Computer Engineering, University of Alberta, Alberta, Canada Systems Research Institute, Polish Academy of Sciences, Warsaw, Poland Marios M. Polycarpou, Department of Electrical and Computer Engineering, KIOS Research Center for Intelligent Systems and Networks, University of Cyprus, Nicosia, Cyprus Imre J. Rudas, Óbuda University, Budapest, Hungary Jun Wang, Department of Computer Science, City University of Hong Kong, Kowloon, Hong Kong

The series “Lecture Notes in Networks and Systems” publishes the latest developments in Networks and Systems—quickly, informally and with high quality. Original research reported in proceedings and post-proceedings represents the core of LNNS. Volumes published in LNNS embrace all aspects and subfields of, as well as new challenges in, Networks and Systems. The series contains proceedings and edited volumes in systems and networks, spanning the areas of Cyber-Physical Systems, Autonomous Systems, Sensor Networks, Control Systems, Energy Systems, Automotive Systems, Biological Systems, Vehicular Networking and Connected Vehicles, Aerospace Systems, Automation, Manufacturing, Smart Grids, Nonlinear Systems, Power Systems, Robotics, Social Systems, Economic Systems and other. Of particular value to both the contributors and the readership are the short publication timeframe and the world-wide distribution and exposure which enable both a wide and rapid dissemination of research output. The series covers the theory, applications, and perspectives on the state of the art and future developments relevant to systems and networks, decision making, control, complex processes and related areas, as embedded in the fields of interdisciplinary and applied sciences, engineering, computer science, physics, economics, social, and life sciences, as well as the paradigms and methodologies behind them. Indexed by SCOPUS, INSPEC, WTI Frankfurt eG, zbMATH, SCImago. All books published in the series are submitted for consideration in Web of Science. For proposals from Asia please contact Aninda Bose ([email protected]).

More information about this series at https://link.springer.com/bookseries/15179

Marek Gzik Zbigniew Paszenda Ewa Piętka Ewaryst Tkacz Krzysztof Milewski Jacek Jurkojć •

•

•

•

•

Editors

Innovations in Biomedical Engineering

123

Editors Marek Gzik Faculty of Biomedical Engineering Silesian University of Technology Zabrze, Poland

Zbigniew Paszenda Faculty of Biomedical Engineering Silesian University of Technology Zabrze, Poland

Ewa Piętka Faculty of Biomedical Engineering Silesian University of Technology Zabrze, Poland

Ewaryst Tkacz Faculty of Biomedical Engineering Silesian University of Technology Zabrze, Poland

Krzysztof Milewski American Heart of Poland S.A. Ustroń, Poland

Jacek Jurkojć Faculty of Biomedical Engineering Silesian University of Technology Zabrze, Poland

ISSN 2367-3370 ISSN 2367-3389 (electronic) Lecture Notes in Networks and Systems ISBN 978-3-030-99111-1 ISBN 978-3-030-99112-8 (eBook) https://doi.org/10.1007/978-3-030-99112-8 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023, corrected publication 2024 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Contents

Experimental Research in Biomedical Engineering Furniture for Elderly People . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Iwona Benek, Iwona Chuchnowska, and Kamil Joszko

3

Thermal Stabiliser of Knee Joint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Iwona Chuchnowska, Katarzyna Białas, Iwona Benek, Zbigniew Opilski, Jadwiga Małecka, Wojciech Michalec, Patryk Mielniczek, Oliwia Nowicka, Paweł Podsiedlik, and Piotr Szaflik

13

MONITORING HELMET—The Use of Thermal Imaging to Monitor the Epidemic Threat Caused by the Corona Virus . . . . . . . . . . . . . . . . . Iwona Chuchnowska, Ewa Lach, Iwona Benek, Maksym Brzęczek, Aleksandra Dziwoki, Michał Kluk, Grzegorz Gruszka, Marek Ples, Michał Kudela, Aleksander Mekail, and Zuzanna Rodak Electronic Stethoscope as a Tool Supporting Physician’s Work During the Covid-19 Pandemic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Iwona Chuchnowska, Katarzyna Białas, Adam Gorol, Mateusz Gutkowski, Piotr Hałacz, Patryk Mielniczek, and Marta Pluta The Epidemiological Analysis of the Atlantoaxial Joint Instability in Children and Youth with Down Syndrome Attending Special Facilities in Upper Silesia (Poland) Based on the Special Olympics Radiological Examination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Iwona Doroniewicz, Paweł Linek, Andrzej Knapik, Adam Posłuszny, Aleksandra Masłowska, and Andrzej Myśliwiec Trends in Development of Balance Dysfunctionalities Rehabilitation Equipment Using Virtual Reality—A Literature Review . . . . . . . . . . . . Grzegorz Gruszka, Piotr Wodarski, Marek Ples, Marta Chmura, Andrzej Bieniek, and Jacek Jurkojć

23

33

41

49

v

vi

Contents

Biomechanical Assessment of Selected Motion and Cognitive Exercises in the “Neuroforma” Rehabilitation System . . . . . . . . . . . . . . . . . . . . . . Agata Guzik-Kopyto, Katarzyna Nowakowska-Lipiec, Piotr Szaflik, Oliwia Nowicka, and Robert Michnik Thermovision-Based Human Body Temperature Measurement Supported by Vision System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Damian Krawczyk, Paulina Sośniak, Weronika Czech, Michał Swierzy, Kajetan Ładoś, Łukasz Seweryn, Michał Zwardoń, Sławomir Suchoń, Wojciech Wolański, Rafał Setlak, and Ziemowit Ostrowski

67

77

Critical Analysis of Recreational Activities as a Method to Reduce Obesity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Małgorzata Matyja, Joanna Szołtysek, and Andrzej W. Mitas

87

Assessment of the Time of Electromechanical Muscle Response to a Given Rhythmic Sound Stimulus . . . . . . . . . . . . . . . . . . . . . . . . . . Robert Michnik, Aneta Danecka, Anna Mańka, and Andrzej W. Mitas

95

Is the Coronavirus Pandemic Going to ‘Kill’ the Physical Activity of Young People? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 Robert Michnik, Katarzyna Nowakowska-Lipiec, Katarzyna Jochymczyk-Woźniak, Aneta Danecka, Karolina Mika, and Hanna Zadoń Therapeutic Use of TMS in Psychiatric Disorders . . . . . . . . . . . . . . . . . 113 Paulina Putko Influence of Back Muscle Activation in Pathological Posture Assessment Based on Thermal Imaging . . . . . . . . . . . . . . . . . . . . . . . . . 119 Patrycja Romaniszyn-Kania, Marta Danch-Wierzchowska, Damian Kania, Daniel Ledwoń, Anna Mańka, Monika Bugdol, Marcin Bugdol, Karol Bibrowicz, Andrzej Myśliwiec, and Andrzej W. Mitas Body Postures During Sitting in Different Positions . . . . . . . . . . . . . . . . 129 Hanna Zadoń, Anna Miller, Katarzyna Nowakowska-Lipiec, Katarzyna Jochymczyk-Woźniak, and Robert Michnik Engineering of Biomaterials Adhesion of Staphylococcus Aureus on Various Biomaterial Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 Marcin Basiaga, Zbigniew Paszenda, Marcin Kaczmarek, Witold Walke, Agata Sambok-Kiełbowicz, Wojciech Kajzer, Anna Taratuta, Julia Lisoń, Magdalena Szindler, and Alicja Kazek-Kęsik Developing the Technology for the Production of Personalized Polylactide Plates for Bone Assemblies Reinforced with Glass Fiber . . . 149 Agnieszka Dubiel, Witold Walke, and Jarosław Żmudzki

Contents

vii

Determination of the Breaking Force of Surgical Threads with the Use of a Testing Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Anita Kajzer, Kamila Kozioł, Wojciech Kajzer, and Halina Malinowska Study of Physical Properties of Additively Manufactured and Postprocessed 3D Porous Structures Intended for Implants . . . . . . . . . . . . . 167 Wojciech Kajzer, Katarzyna Gieracka, Mateusz Pawlik, Marcin Kaczmarek, and Anita Kajzer Impact of 3D Printing Materials on Bone Phantom Features . . . . . . . . . 179 Marta Kiel-Jamrozik, Wojciech Jamrozik, Mateusz Pawlik, and Jakub Goczyla Porous Structure and Surface Chemistry of Biomorphous Composite Derived from Carbonized Yucca Covered by Thin Film of Chitosan and Its Application to Remove Hazardous Substances . . . . . . . . . . . . . . 187 Justyna Majewska, Marta Krzesińska, Temenuzhka Budinova, Anna Filipowska, Nartzislav Petrov, and Boyko Tsyntsarski Novel Hemocompatible Thiol-yne Based Photopolymers Obtained by the Advanced Stereolithography (SLA) Processing with Strongly Improved Surface Smoothness by a Novel Exposition Approach for Anti-aliasing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Roman Major, Marcin Surmiak, Maciej Gawlikowski, Romana Schwarz, Marcin Kot, Justyna Wiecek, and Juergen M. Lackner Evaluation of the Physicochemical Properties of Passive Layers Produced on NiTi Alloys for Use in the Cardiovascular System . . . . . . 217 Anna Taratuta, Zbigniew Paszenda, Marcin Basiaga, Magdalena Antonowicz, Witold Walke, and Damian Nakonieczny Research on the Influence of Anodic Oxidation Parameters on the Corrosion Resistance of Titanium Alloys . . . . . . . . . . . . . . . . . . . 225 Karolina Wilk and Janusz Szewczenko Flexural Modulus of Synthetic Femur Polyurethane Foam Bone Components Based on Three-Point Bending Tests . . . . . . . . . . . . . . . . . 233 Krzysztof Zerdzicki Informatics and Modelling in Biomedical Engineering The Application for Reading Comprehension and Reading Speed Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 Łukasz Grabny, Rafał Doniec, Szymon Sieciński, Natalia Piaseczna, and Konrad Duraj

viii

Contents

Numerical Investigations of Mechanical Properties of Head Protection Systems Against the Effects of Dynamic Loads . . . . . . . . . . . . . . . . . . . 255 Aleksandra Jędrzejewska, Kamila Wiśniewska, Monika Ratajczak, and Tomasz Klekiel Influence of the Reorientation Function on Brodmann Areas Detection Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 Ilona Karpiel Comparison and Evaluation of Models for Predicting Immunogenicity of Viral Antigens of the pMHC Complex from Murine Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Gracjan Kątek, Marta Gackowska, Karol Harwtig, and Anna Marciniak Finite Element Analysis of Lumbar Disc Implant, in Aspect of Treating Degenerative Changes in Spine . . . . . . . . . . . . . . . . . . . . . . 291 Dawid Kęszycki, Bogdan Dybała, Grzegorz Ziółkowski, and Patrycja Szymczyk-Ziółkowska Controlling of the Upper Limb Prosthesis Using Camera and Artificial Neural Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301 Agata Mrozek, Martyna Sopa, Jakub K. Grabski, and Tomasz Walczak Lateral Tibial Condyle Fracture Stabilization—A Numerical Analysis . . . 311 Olimpia Promirska and Jakub Słowiński Extreme Compression of the Electrocardiographic Signals Using Matching Persuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 Sandra Śmigiel Experimental Study the Blood Flows in a Transparent Models of a Blood Vessels with Bifurcation—Preliminary Report . . . . . . . . . . . . . . . 335 Wojciech Wolański, Marek Ples, Marta Sobkowiak-Pilorz, Grzegorz Gruszka, Michał Burkacki, Sławomir Suchoń, and Marek Gzik ARM-200 - Upper Limb Rehabilitation Robot . . . . . . . . . . . . . . . . . . . . 341 Andrzej Michnik, Mariusz Sobiech, Jakub Wołoszyn, Mirella Urzeniczok, Aleksander Sobotnicki, Rafał Kowolik, and Krzysztof Cygoń Correction to: Porous Structure and Surface Chemistry of Biomorphous Composite Derived from Carbonized Yucca Covered by Thin Film of Chitosan and Its Application to Remove Hazardous Substances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Justyna Majewska, Marta Krzesińska, Temenuzhka Budinova, Anna Filipowska, Nartzislav Petrov, and Boyko Tsyntsarski

C1

Experimental Research in Biomedical Engineering

Experimental research has a very wide range of applications in the field of biomedical engineering, as evidenced by the publication topics of this chapter. It is a key part of each research, enabling the verification of theses as well as indicating the possibilities of implementing the proposed practical solutions. The chapter presents works covering the subject of biomechanical measurements and practical verification of the designed devices. muscle response to and given rhythmic sound stimulus, the impact of the pandemic on physical activity, body posture while sitting and the possibility of using virtual reality in rehabilitation of the ability to maintain balance are examples of works included in the chapter. Additionally some practical solutions, like innovative furniture for the elderly or an electronic stethoscope are also presented.

Furniture for Elderly People Iwona Benek, Iwona Chuchnowska, and Kamil Joszko

Abstract The article presents a description of the results of interdisciplinary cooperation. It is related to the implementation of research carried out in the form of Project Based Learning as a part of the project ‘Silesian University of Technology as the Centre for Modern Education based on research and innovation’. Employees and students of the Faculty of Biomedical Engineering and the Faculty of Architecture of the Silesian University of Technology have attempted to design some furniture for senior integrated with control systems located in the LeonardoLab—2014a room for testing technological solutions for the elderly. Keywords Elderly dependents · LeonardoLab

1 Introduction The issue of designing for older people becomes an important question in connection with demographic changes and an increasing number of people who need support in everyday life due to their advanced age and health problems. For a number of years, this topic has been developed at the Faculty of Architecture and at the Faculty of Biomedical Engineering of the Silesian University of Technology in Gliwice. In this work, the issues related is caring for a dependent elderly person were considered. Based on this project, the need to determine optimal design solutions related to

I. Benek Faculty of Architecture, Silesian University of Technology, Gliwice, Poland e-mail: [email protected] I. Chuchnowska (B) · K. Joszko Faculty of Biomedical Engineering, Silesian University of Technology, Gliwice, Poland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_1

3

4

I. Benek et al.

shaping the internal environment has been identified. Important research issues have also emerged and what is more, they should be focused on shaping elements supporting the care of dependent people in order to be used by them in everyday activities.

1.1 Purpose of the Work The main purpose of this work is to determine the optimal—functional, spatial and formal features of furniture in the room for an elderly dependent person. Items of furniture ought to support the treatment process. They should create proper conditions, which supports the functional competences of older patients. The research part focuses on the stay zone of a dependent elderly person and analyzes some ergonomic requirements and psychological needs resulting from their state of health. The following assumptions were made in this work: – adequate spatial conditions are one of the most important factors supporting older patients in their current functional competences, – individual elements of the environment for an elderly dependent person should increase their activity, mobility and independence, – the basic criterion in designing the space surrounding an elderly person is the reduction of stress resulting from maintaining the proper level of control over the environment. A very important source of data and the basis for formulating conclusions about the comfort felt by an elderly dependent person and working conditions of caregivers is information obtained from users about the ways of using facilities in the Geriatric wards [1].

1.2 Methodology At the initial stage of work the methods of analysis and criticism of the literature were used, thanks to which the goals of the project were defined. The main trends in equipping rooms related to the stay of a dependent elderly person were indicated. Ergonomic guidelines determined a reference point for the functional and formal conditions in shaping the pieces of furniture in rooms dedicated to the elderly [3]. The work also included pilot studies at the Geriatric Hospital in Katowice, this action helped to characterize the surrounding environment of the geriatric patients and verify the proper design assumptions. Research techniques were applied in the form of: interviews with medical and nursing staff, observation of the functioning of the ward, tracking the patient’s path and interviews with selected patients. The research was carried out for separate functional, behavioral and technical categories. Observation of the daily habits of dependent people made it possible to notice the frequent use of bedside cabinets and hospital cabinets. It proved the primary role that proper furniture plays in organizing the control centre of an elderly person. Practical access to everyday objects, and carrying our independent daily activities are the basis for a sense of security and control for dependent people. Hospital patients valued cabinets by their beds and they also reported many disadvantages. The remarks concerned: the

Furniture for Elderly People

5

size and availability of the countertop (too small a surface, no possibility of height adjustment), drawers (not accessible to a person who is lying down), no possibility of recharging the phone, as well as being able to putting away everyday objects in a safe way (medicines, glasses, etc.). A list of the most frequently used items and devices was prepared during the observations (for instance: a cabinet, a lamp, a contact plug, a telephone, glasses, newspapers, an alarm clock, a wrist watch, a calendar, a comb, a cosmetic bag, photo frames, a mascot, a medicine pack and a cup, etc.). The conclusions from the tests have been verified using the method of study in design-through conceptual work for selected pieces of equipment, it allowed the identification of the most important activities in creating an environment supporting older patients. The result of the research conducted is the formulation of design guidelines in the form of a furniture design (a bedside cabinet) serving a dependent elderly person. The project site was the LeonardoLab-a smart laboratory located at the Faculty of Biomedical Engineering at the Silesian University of Technology in Zabrze. From the technological aspect, the LeonardoLab is operated by the NAZCA building automation system designed by the initiator of the project, the APA Group in Gliwice. The aim of the LeonardoLab is to develop technological solutions that will make it easier for medical staff to care for the elderly and the injured. It is used to test prototypes of products, services and technologies.

2 Analysis of the Problem In the report entitled The Situation of Caregivers of Elderly Dependents [1], it is written: Seniors—it is a very wide term, depending on the definition, people from 55 to 65 and above are included in this group. (...) Older dependents are also a heterogeneous group. First of all, the causes of dependence are different, and there are also various areas in which dependent seniors need support. Designing an environment for the elderly requires a detailed and insightful approach that includes the ability to adapt it on a regular basis to the changing skills and needs of seniors. It is also necessary to take into account the disabilities commonly encountered like motor, auditory, visual and mental types. Susanne Iwarsson (professor of rehabilitation at the University of Lund, Sweden) in the book Housing Enabler [2], gives a set of the following disabilities that impede independence: – difficulties in interpreting information (dementia, Alzheimer’s disease), – partial or total loss of vision, – partial or total loss of hearing, – problems with keeping balance, – lack of physical coordination, – poor physical condition, – difficulties in moving the head, – difficulties in raising the arms, – difficulties in lifting and catching objects, – loss of superior manual activities, – difficulties with bending down and kneeling, – walking on crutches, – using a wheelchair, – being overweight (a rare situation among the very elderly) or being extremely tall. To sum up, the ergonomic requirements of older people are specific due to the various physical dysfunctions. The disabilities are associated with difficulties in moving around and using crutches, walking sticks, walkers or wheelchairs.

6

I. Benek et al.

2.1 Overview of the Solutions In order to be able to specify the conceptual assumptions, the components of the equipment and the support for the elderly dependents were reviewed. One of the groups of items were cabinets and bedside tables. The analysis concerned material, colour and functional solutions. These products were characterized by a lack of adaptation to the needs of elderly dependents in terms of ergonomics (no possibility to use the worktop in a lying position, no access to drawers), in forms of personal style preferences terms (for instance inadequate colours) and lack of electronic systems. Another review was focused on a group of electronically controlled items. Examples of such the furniture was inspired by the control systems available in the LeonardoLab. Well-developed telecare services were also analyzed such as the OPOS24 telecare unit, which uses communication via GSM network (the set is equipped with a gas leak sensor, a motion sensor, an open door sensor and an (SOS) band or keychain [4]), and mobile telecare systems (e.g. PillDrill—the device sends notifications to the smartphone, in the form of a message. The device scans the pillbox, then checks whether the appropriate drug has been selected and opens the lid [5]).

3 Project 3.1 Conceptual Work The next step was to analyze the interior design for dependent elderly people in terms of design, rehabilitation and implementation of modern technologies.Three proposals for furniture were put forward—a bed, a cabinet and an armchair. Each proposed piece of furniture is often used by the elderly or handicapped dependent. A bed has already been placed in the LeonardoLab, which has been attached to the control system in an intelligent room. The NAZCA system controls the built-in actuators in the bed, it also allows you to control the bed using an additional optical interface, in which the camera recognizes the facial expressions of the person lying on the bed and uses a mobile application on the smartphone. The Android system allows you to control modules plugged into NAZCA through voice commands. The advanced bed control features inspired the team to construct more furniture for the LeonardoLab studio. Additionally, further work was focused on the bedside cabinet design because it was universal in terms of functionality. Four variants of the bedside table were proposed, what is more, it was important to think about the available functional solutions that could be implemented by an interdisciplinary team of bioengineers and architects. The first option Fig. 1 was based on designing a movable column at the cabinet, which would raise the drawer and counter to the appropriate height, adapted to the user’s hands range.

Furniture for Elderly People

7

Fig. 1 The model of the bedside cabinet with a presentation of functional in one of the material solution Source Own data

Fig. 2 The first option based on designing a movable column at the cabinet Source Own data

In the second variant Fig. 2, mobile cabinets moved horizontally, which allowed the patient to adjust the storage area to their changing needs. The third version Fig. 3, consisting of a table with a top, offered the possibility of controlling its height. The fourth type of bedside cabinet was also mobile. In this option, the drawers could be hidden in the carrying body of this furniture (Fig. 4). The further parts of the project were focused on the third solution (the cabinet with raised countertop).

8

I. Benek et al.

Fig. 3 The second variant—mobile cabinets moved horizontally Source Own data Fig. 4 The third version consisting of a table with a raised top Source Own data

3.2 The Evolution of the Project After analysis, it was decided that the designed piece of furniture would combine the features of a bedside table and a cabinet with drawers.This solution is the most secure, thanks to which the product will compete with the latest technologies appearing on the market over a long period of time. It will be clearly more expensive than other proposals, which despite the lowest level of advancement is characterized by the highest mobility. The most important function of the furniture is the possibility to adjust the height of the countertop. The aluminum profile from Bosch Rexroth was chosen as the basis of the construction because of its durability and simplicity. The first version of the project (Fig. 5) took into account the implementation of one electric column, which allows the table to fit enter under hospital bed. This solution was characterized by a very massive column, which took up a lot of usable space. This version was disqualified due to the very limited range of the column extension and its large dimensions. The next stages of project involved using a smaller lift system—the ram. This version of the project assumed the construction of two table tops, one fixed and the other one lifted and rotated. Lifting only one table top resulted in a large difference in height between the work planes. In addition, the seemingly small ram required a proper construction to improve the stability of the structure,

Furniture for Elderly People

9

Fig. 5 Functional model of furniture. The first version of the project (on the left). The third version of the project (on the right) Source Own data

and thus the size of the table increased significantly. In the third version, two electric columns were used to raise the countertop (Fig. 5). The rotating countertop was connected to the main table thanks to a swivel bearing and made it possible to easily rotate it. The space for the drawers was placed right under the top, thus creating a free space at the bottom of the cabinet. In the final version of the project, the columns have a very large range of motion, furthermore, they are extremely stable, and have a special protection to prevent objects from being destroyed. The rotating countertop is equipped with an additional folding part, thanks to that the worktop is accessible in a lying position (Fig. 6). According to the original assumption, the aluminum profile from Bosch remained the basis of the structure. The drawers are mounted on solid roller guides. The selection of electronically controlled functions that can be implemented in the furniture was proposed, for instance to automatically adjust the cabinet’s height using a drive system of Jiecang’s desk. The system is based on a dual synchronized electric drive with Hall effect sensors. The system is controlled by a remote control provided by the manufacturer. For the purpose of the project, the control function was provided by the NodeMCU ESP8266 board, which was chosen due to its integrated Wi-Fi wireless communication module, due to which it was possible to communicate with the NAZCA system, responsible for controlling LeonardoLab (Fig. 7). The communication protocol between the NodeMCU board and NAZCA system is Modbus TCP/IP. Work related to the implementation of the functional model was carried out simultaneously with the development of its design (Fig. 6). The aspect of functionality of the bedside cabinet and the possibilities of modifying the furniture by adjusting it to the individual needs of the user were also important for the team. The final effect is a prototype of furniture for seniors, which can be integrated with the NAZCA systems found in LeonardoLab (Fig. 7). In this way, the task of designing an innovative component of the LeonardoLab equipment was successfully completed in a comprehensive and interdisciplinary way.

10

I. Benek et al.

Fig. 6 The model of the bedside cabinet with a presentation of functional in one of the material solution Source Own data

Fig. 7 The model of the bedside cabinet with a presentation of functional in one of the material solution Source Own data

4 Conclusion As a part of the project, the conducted interdisciplinary research was conducted with the participation of the Faculty of Architecture, the Faculty of Biomedical Engineering, the APA Group—a company that deals with automated technology and systems for intelligent construction, EMC Silesia Ltd.—a Geriatric Hospital with the Research and Development Centre. One of the aims of the project was to give it an overall aspect of interdisciplinarity basis. Undoubted advantages of such cooperation include a comprehensive recognition of the needs of elderly dependents, better results of the preparation of the functional model, while paying attention to formal solutions. The biggest difficult, was to determine the needs and limitations at the stage of design. On the one hand, it generates some problems, but on the other hand it gives the opportunity to learn more about solving them. Eventually, such cooperation gives better results at the design and implementation process—people from various fields are able to assess both the progress of a given project and the final effect, as well as the further possibilities of its development or potential corrections

Furniture for Elderly People

11

Acknowledgements This research was supported by the European Union from the European Social Fund in the framework of the project "Silesian University of Technology as a Center of Modern Education based on research and innovation” POWR.03.05.00-00-Z098/17.

References ´ ˛ska entitled The situation of carers of 1. Benek I (2018) Report on direct consultations in Ruda Sla ´ ˛ska (PL) dependent elderly people, Fundacja Rzecz Społeczna, Ruda Sla 2. Iwarsson S, Slaug B (2001) Housing enabler. An instrument for assessing and analysing accessibility problems in housing. Studentlitteratur, Lund 3. Niezabitowska ED, Szewczenko A, Benek I (2017) Needs of elderly people in facilities with a care function. Design guidelines. Publishing House of the Silesian University of Technology, Gliwice (PL) 4. http://www.opos24.pl/nasze-urzadzenia/ (28.09.2020) 5. https://www.pilldrill.com (28.10.2020)

Thermal Stabiliser of Knee Joint Iwona Chuchnowska, Katarzyna Białas, Iwona Benek, Zbigniew Opilski, Jadwiga Małecka, Wojciech Michalec, Patryk Mielniczek, Oliwia Nowicka, Paweł Podsiedlik, and Piotr Szaflik

Abstract Cryotherapy, also known as cold therapy, is a method consisting in the lowering of the temperature of tissues using water, ice, air, ethyl chloride, dinitrogen monoxide or liquid nitrogen. Treatment with low temperature may be applied generally or locally. General cryotherapy consists in the cooling of the whole body, whereas local one refers to the decrease in the temperature of skin and tissues. In this method, the cooling area should not be too big. The research project involved the second type of the above-mentioned cooling methods. The project entitled “Thermal Stabiliser of the Knee Joint” was implemented by a team of students. The objective of this project was to develop and then fabricate a cooling system using the Peltier module, including the attachment of the system to the stabiliser of the knee joint. The developed device will find application in local cryotherapy of the knee joint. This type of therapy is commonly applied in early post-trauma conditions with the damage to soft tissues (spraining and dislocation of joints, contusions of muscles), chronic inflammatory and degenerative diseases as well as overload disorders (degenerative diseases of knee joints, rheumatoid arthritis of joints), conditions of increased muscle tension and limitation of joint mobility, chronic oedema and joint exudates. The first stage of the research encompassed a review of the existing solutions of ortheses and methods of cooling the knee joint. Next, two concepts of the device were developed and their visualization was prepared in the Inventor software programme. The third phase involved the development of the documentation concerning the project I. Chuchnowska (B) · O. Nowicka · P. Szaflik Faculty of Biomedical Engineering, Silesian University of Technology, 41-800 Zabrze, Poland e-mail: [email protected] K. Białas · W. Michalec · P. Mielniczek · P. Podsiedlik Faculty of Mechanical Engineering, Silesian University of Technology, 44-100 Gliwice, Poland I. Benek · J. Małecka Faculty of Architecture, Silesian University of Technology, 44-100 Gliwice, Poland Z. Opilski Faculty of Electrical Engineering, Silesian University of Technology, 44-100 Gliwice, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_2

13

14

I. Chuchnowska et al.

implementation and the applied technology. This was done using the CAD software programme. The project was also analysed with a view to satisfying the functional requirements. In addition to that, necessary elements were printed using a 3D printer. The last stage involved the manufacturing of a prototype of a functional device taking into consideration electronics and an Android system application. Keywords Cryotherapy · Local cryotherapy · Orthesis · Knee joint stabiliser · Knee joint

1 Introduction Cryotherapy involves activities aiming to alleviate pain by means of various types of substances having cooling properties. Such substances include ethyl chloride, dinitrogen monoxide, liquid nitrogen, water or ice. The birth of cryotherapy for the purpose of pain alleviation goes back to the 20th century, when a medium in the form of cold water or ice was laid on the body organ affected by disease. The rule of the operation of cryotherapeutical procedures remains the same, however, the technological progress contributed to the development of new, more effective, methods of cold therapy. With the passing of time, the discovery of further substances with cooling properties led to the distinguishment of two types of cryogenic procedures, namely local and general [1]. Local cryotherapy consists in the placement of a cooling compress directly on the sites of pain occurrence. This method aims to lower the temperature of skin and tissues by laying there, for instance plastic bags filled with specialist cooling gel, water of lowered temperature or ice. It should be noted that a local cryotherapy procedure should not exceed 20 min due to the fact that a longer period of time may cause adverse effects in the form of frostbite. The procedure lasting 20 min is sufficient to cool the tissues down to a temperature of 13 ◦ C. This is an optimum temperature to obtain a desired effect of the loss of sensation and analgesia [2, 3]. Whole body cryotherapy is a method of rehabilitation involving the cooling of the patient’s whole body by staying in a special cryogenic chamber which is filled, for instance, with nitrogen. Such treatment lasts no longer than 3 min due to very low temperatures, dropping even to –100 ◦ C. The advantages of this type of treatment include the improvement of psychological condition, relaxation of muscles and disappearance of internal pains in the organism [1–3].

2 Materials and Methods Tests of power consumption of the device were conducted using an ammeter, a voltmeter and a thermal-imaging camera Flir One Pro, whose measurement range falls between –20–400 ◦ C. The measurements were carried out three times and the obtained results were subjected to the calculation of the arithmetic mean. The perfor-

Thermal Stabiliser of Knee Joint

15

mance of measurements in the above-described way aimed to minimize as much as possible the range of error of each device. The design of the casings of two separate devices was developed in the CAD Autodesk Inventor 2020 software programme. It was done taking into consideration optimum dimensions. It is possible to ‘seal’ the casing model with screw joints for the purpose of using hydraulic and electric devices. During the preparation of models for 3D printing, the Z-suite software programme was used, whereas 3D printing was executed with the application of a Zortrax m200 printer. The appearance of the whole object and the way of attachment of all elements on the user’s body were designed in the Photoshop and SketchUp software programmes. The software for controlling the device was developed in the programming environment of the Android Studio, which enables the creation of applications in a compatible way with the majority of mobile phone systems based on the Android system. A three-dimensional model of the designed device and the technical documentation of all the elements which were necessary for its creation were prepared in the Autodesk Inventor 2020 software programme.

3 Results of Investigations The conducted investigations refer in particular to the obtained temperatures, cooling time and power input necessary to supply power to the whole mechanism. A small bag filled with liquid was placed on the knee joint in such a way as to effectively lower the temperature of the site where the patient may feel pain. The measurement of the temperature of the bag containing liquid was conducted using a thermal-imaging camera (Fig. 1).

Fig. 1 Temperature of liquid a at the beginning of measurements b after 15 min of tests b after 40 min of tests

16

I. Chuchnowska et al.

Apart from the temperature, the measurements involved also the power consumption of the whole device. The results were averaged in order to minimize the measurement error of devices (Table 1, Fig. 2). A belt, playing a function of a holder for the casing, was fitted on the hips. The casing houses, among other things, a battery supplying the whole mechanism with power and electronic control elements. The battery is connected to another casing by means of supply cables. The second casing is fitted on the thigh slightly above the knee orthesis and contains elements of heat removal (heatsink) which cool the

Table 1 Results obtained during the measurements of temperature Time (min) Electric Temp. (◦ C) Time (min) Electric current (A) current (A) 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28

5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5

22,5 22,8 23,5 24,5 23,9 22,2 22,1 21,8 20,9 20,4 19,9 19,5 19,6 19,2 18,9

30 32 34 36 38 40 42 44 46 48 50 52 54 56 58

5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 5,5 0,6

Temp. (◦ C) 18,7 18,2 18 15,1 15,0 14,5 13,0 12,8 12,1 11,6 10,9 10,5 10,2 9,7 9,0

Fig. 2 Diagram of the dependence of the obtained temperature on the duration of the measurements

Thermal Stabiliser of Knee Joint

17

Fig. 3 Visualization of the developed design of the device

substance in the plastic bag. A water pump is responsible for the transport of liquid. The pump forces the substance cooled by the Peltier module into the bag and receives the liquid warmed up by the knee, creating thus a closed cycle. The whole system is stabilised by the knee orthesis, which stiffens the joint and provides protection during the rehabilitation of the knee joint (Fig. 3). One of the assumptions of the project was to apply a microcontroller of the Arduino family. The selected plate: Arduino Nano 33 BLE is characterized by very small dimensions and extremely low power consumption, amounting to less than 20 mA. Module HC-06 is in charge of the connection of a Bluetooth microcontroller and a mobile phone application with the Android system. The communication is maintained by means of a series connection in the Bluetooth 2.0 standard. The obtained temperature is measured with analogue temperature sensor LM35. The selected sensor has a wide range of temperature readings (from –55 to 150 ◦ C) with accuracy of 0.5 ◦ C. The executive element is a 60 W Peltier module. In order to determine the power needed for the cooling of the system, it was necessary to take into consideration all elements which charge electric energy from the supply system. In this case, the elements were as follows: Peltier module, cooling fan, pump of liquid and electronic control system. It was assumed that power reserve should be also taken into consideration. It was decided that a maximum end power at one Peltier module should equal 65 W. The device is fed by a pack of lithium-ion batteries. The programme, which is implemented into the Arduino microcontroller, is secured against setting too low temperatures. The default lowest temperature is 0 ◦ C. The programme

18

I. Chuchnowska et al.

Fig. 4 Diagram of the system

also has a safety function turning off the cooling after a certain time. The default maximum time amounts to 3 h (Fig. 4).

4 Android System Application One of the assumptions of the design of a thermal orthesis of the knee joint was the development of an application in the Android system, which would serve the purpose of the management and controlling of the device. The application should be wirelessly connected to the device (for instance, through the Bluetooth interface) and make it possible to start the operation of the device, turn it off and change the set temperature. The interface should be readable and easy to use. It should enable the user to check the current state of the battery and temperature readouts. In order to implement the above-mentioned assumptions, the Bluetooth interface was applied. The main reason for the application of this type of a communication system is the availability of modules - each updated mobile phone with the Android system has the Bluetooth module. In order to complete the final version of the application, the official software programme Android Studio, provided by Google, was used. The fabricated interface can be seen below (Fig. 5). In the central part of the application, there is a chief control button START or STOP, serving the purpose of turning on and off the device. Above the control button, there is a slider which enables setting the target

Thermal Stabiliser of Knee Joint

19

Fig. 5 Interface of the application a in use b ready to use

temperature of the device (current and target temperatures are shown in the upper part of the application on the right and left side, respectively). In the lower part of the application, below the chief control button, there is an instruction field informing the user about the operating status and purpose of the application. In addition, the screen of the application enables the readout of the current temperature of the device and the battery status. Following the start-up, the programme makes a series of attempts to connect with the device by means of a wireless BT communication. When the connection is not possible, the user receives a message about an error and is asked to make sure if the device has been turned on. In order to repeat the communication attempt, it is necessary to turn on the application once again. When the device gets connected to the application, the user is informed about that and asked to select temperature and press the START button. The programme sends a ‘start’ order and a set temperature to the device.

5 Discussion The conducted tests show the cooling efficiency based on one Peltier’s module. Calculations performed by means of analytical methods provided the basis for the estimation of safe capacity of a battery being able to supply power to the whole device. The above-mentioned capacity amounts to 10 Ah. Taking into consideration low cooling efficiency of the Peltier module in relation to significant power input, it would be advisable to improve the device. One of the methods of efficiency improvement is the addition of a subsequent cooling module to the whole system, which would decrease the time of cooling the liquid to a required temperature down to 17 min of the device operation. However, this is connected with the application of a bigger battery of greater capacity. As a result, a heavier device, which is held and attached to the hips (Fig. 6), will cause the loss of comfort by the user. Another way of the

20

I. Chuchnowska et al.

Fig. 6 Fabricated cooling system attached to the knee joint stabiliser

system improvement is the storage of the bag with the cooling substance, prior to its use, in the refrigerator until it achieves a suitable temperature for the analgesic effect. The role of the device mechanism would be then just maintaining the already lowered temperature on the site of pain occurrence. Such a method would contribute to the functionality and comfort of use in patients. The obtained temperature results are burdened with a high tolerance error. This fact results from imperfections of the thermal-imaging camera, whose long calibration time and fluctuations of displayed temperatures are far from real-life results.

6 Summary The implemented project aiming at the local cryotherapy of the knee joint with the simultaneous use of the stabiliser successfully underwent the investigations. Further development of the above-described cryogenic system is planned in order to improve its efficiency and design as well as increase the comfort of use for the patients. All this will enable the adaptation of the system to the expected results. In spite of numerous problems in different phases of the construction, the creation of the mechanism has enabled better understanding of physiotherapeutical issues and broadened the knowledge of local cryotherapy used in the knee joint rehabilitation. Acknowledgements This research was supported by the European Union from the European Social Fund in the framework of the project “Silesian University of Technology as a Center of Modern Education based on research and innovation” POWR.03.05.00- 00-Z098/17.

Thermal Stabiliser of Knee Joint

21

References 1. Kalinowska K (2017) Leczenie zimnem w praktyce fizjoterapeuty. Tom XLV, Kujawska Szkoła Wy˙zsza we Włocławku 2. Straburzy´nska-Lupa A, Straburzy´nski G, Straburzy´nska-Migaj E (2008) Fizjoterapia z elementami klinicznymi, Tom 1, Warszawa 3. Chorowski M (2012) Zastosowanie niskich temperatur w biomedycynie. Wprowadzenie do kriogeniki, Rozdział I, Wrocław

MONITORING HELMET—The Use of Thermal Imaging to Monitor the Epidemic Threat Caused by the Corona Virus Iwona Chuchnowska, Ewa Lach, Iwona Benek, Maksym Brze˛czek, Aleksandra Dziwoki, Michał Kluk, Grzegorz Gruszka, Marek Ples, Michał Kudela, Aleksander Mekail, and Zuzanna Rodak

Abstract The objective of an interdisciplinary team of IT specialists, bio-engineers, architects, a specialist in thermovisual measurements and medical personnel was to develop an apparatus enabling a remote measurement of human temperature using a thermal imaging camera coupled with a mobile phone and the Augmented Reality technology. The team designed a portable device which makes it possible to conduct measurements in an automatic way without the use of hands. Keywords Thermovision · Covid-19 · Monitoring · Thermal imaging

1 Introduction A coronavirus identified in 2019, SARS-CoV-2, has caused a pandemic of respiratory illness, called COVID-19. COVID-19 is an infectious disease that can be severe and has caused millions of deaths around the world as well as lasting health problems in some who have survived the illness. The coronavirus can be spread from person to person through droplets and virus particles released into the air when an infected person breathes, talks, laughs, sings, coughs, or sneezes. Larger droplets may fall to the ground in a few seconds, but tiny infectious particles can linger in the air and accumulate in indoor places, especially where many people are gathered and there is poor ventilation. The most common symptoms of COVID-19 are fever, cough, muscle pain, and fatigue. Fever is the symptom the easiest to identify. Since the virus I. Chuchnowska (B) · M. Kluk · G. Gruszka · M. Ples Faculty of Biomedical Engineering, Silesian University of Technology, Gliwice, Poland e-mail: [email protected] E. Lach · M. Brze˛czek · M. Kudela · A. Mekail · Z. Rodak Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, Gliwice, Poland I. Benek · A. Dziwoki Faculty of Architecture, Silesian University of Technology, Gliwice, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_3

23

24

I. Chuchnowska et al.

outbreak, thermal screening using infrared thermometers and thermal imaging has been used at public places to check the body temperature to identify infected among the crowd [5, 6]. This study proposes the design of a simple, cheap, and easy to obtain monitoring helmet that has the capability to detect humans with a fever automatically from the thermal images. The thermal camera technology is combined with AR (Augmented Reality) technology with the use of a standard mobile phone. Additionally, the proposed system is equipped with facial recognition technology. It can automatically measure the temperature of surrounding people and inform if the recorded temperature exceeds the set threshold. This proposed project can be quickly, simply, and cheaply implemented by the health care system and other services that exposed their employees to potentially sick people and could potentially help prevent the wider spread of the coronavirus.

2 Temperature Measurement Two approaches can be distinguished among temperature measurement methods, namely contact and contactless methods of temperature detection. The abovementioned techniques make use of various devices enabling the estimation of the body temperature values in invasive or non-invasive ways. The tools which are applied clinically to measure the body temperature can be divided into contact and contactless devices. Contact devices are devices whose measurement element comes into contact with the tested surface. Such devices are widely used at home and in medical practice. They include liquid and electronic thermometers. Liquid thermometers use that the volume of the liquid in a container grows along with an increase in temperature. he disadvantage of this type of device is its vulnerability to mechanical damage, as well as a relatively long time of measurement lasting from 4 to 10 min. Such a thermometer makes it possible to measure temperature in the oral cavity, armpit or anus. Electronic thermometers most often use thermistors, i.e. semiconductor resistors, whose resistance is strongly dependent on temperature. This type of thermometer is used to measure temperature in the oral cavity, anus and armpits. A great advantage of this type of device is a very short, in relation to liquid and phasechange thermometers, time of measurement, from 10 to 15 s. Unfortunately, contact electronic thermometers also have some disadvantages, namely they are sensitive to electromagnetic fields and require the use of batteries as their means of supply. Contactless devices are devices whose measuring element does not come into contact with the tested surface and the measurement of temperature is conducted at some distance from the tested object. An example of a thermometer based on the above-mentioned contactless operation is an electronic radiation thermometer, which is used both at home and in medical practice. Infrared thermometers detect infrared radiation emitted by each body having its temperature higher than absolute zero. The devices of this type are built using optical detectors, often thermostats,

MONITORING HELMET ...

25

which enable the detection of radiation. The result is displayed just after a few seconds, which makes this method the fastest of all. Unfortunately, a disadvantage of contactless thermometers consists in its lower accuracy in comparison with contact devices, which in most cases equals ±0,2 ◦ C. In addition, factors related to the measurement site (both forehead and ear canal), such as make-up, sweat, hair and dirt, may all generate read-out errors [6]. What is more, due to the fact of using different algorithms, radiation thermometers may show some discrepancy in the obtained results and thus require repeat calibration after a long period of application. Infrared thermometers and thermal imagers basically have the same function. An infrared thermometer uses short-wavelength infrared light to illuminate the area of interest and measures the average surface temperature of a given “spot”. Spots vary in terms of size which depends on the specification of the thermometer and their distance from the target. A thermal imager can provide a more detailed and sophisticated result. It uses mid- or long-wavelength infrared energy and a single image can provide hundreds or thousands of individual temperature readings, one for every pixel view. These images (Fig. 1) can be recorded, and processed for viewing using specialized software. Thermal imaging, which finds application in many areas of life, is a very useful tool for the detection of changes in temperature on the object’s surface. Thermography has broad application in different fields of medicine. The development of technology, first of all, the standardisation of thermographic tests contributed to the growth of popularity of thermovision in medical circles. Thermal imaging makes it possible to carry out and support medical diagnostics in a non-invasive way and, what is more, painless for the patient. In addition, according to the subject literature [3], the thermovision enables the assessment of the efficiency of prescribed medicines and assistance in various types of medical procedures, including cardiovascular procedures or physiotherapeutic treatment [1, 4]. Thanks to the detection of infrared radiation, thermal imaging enables the localisation of the changes of temperature on the surface of the patient’s body, which makes it possible to discover pathological phenomena occurring in the patient’s body. Thermal imaging examination is effective due to the fact that the pathological tissue is characterised by increased blood supply and metabolism, which in turn leads to the higher than average temperature of the skin in an affected area [2]. At present, the above-mentioned technique is used by specialists in various medical fields, from dentistry, ophthalmology, laryngology, rheumatology, etc., to cardiology. The most frequent application of thermal imaging is for the evaluation of the extent of inflammatory condition in joints, e.g. temporomandibular joint, disorders of blood supply, chronic diseases, including neoplasms, most often breast neoplasms. In laryngology, this method enables the diagnosis of the inflammation of nasal sinuses, whereas in endocrinology the diagnosis of thyroid gland diseases. What is important and what should be emphasised is that such an examination should not be the only basis for the diagnosis, but should be treated as an auxiliary test accompanying other examinations, for instance histopathological ones. A thermal imaging camera is a device allowing the recording and visualisation of the temperature distribution on the object’s surface.

26

I. Chuchnowska et al.

3 Materials and Research Methodology The Authors of this work aimed to develop a device which would make it possible to diagnose heightened body temperature in the surroundings in a quick and mobile way. The key factor contributing to the efficiency of this solution is the possibility of use in the way which would physically limit the user to a minimum degree. The requirements of the device have been defined and are as follows: – Automatic detection of individuals located in the user’s surroundings and the recording of their temperature. – Display of the results of temperature measurements in such a way that distracts the user’s attention to a minimum degree. – Ergonomics of the device use. Selection of the mounting system and portable elements which physically limit the user to a minimum degree and ensure flexibility of the mounting. – Affordable price, easy access to the device components. A meeting with an expert allowed the Authors to determine the best places for thermal imaging in humans. During the analysis it was determined that the measurement tests should be carried out within the face of the subject. The temperature on the face can be equal to or lower than the internal temperature if the skin has not been heated from the outside. Therefore, the end result is the warmest point on the face. The work was undertaken in order to determine possible methods of taking automatic measurements of temperature. Thermovisual solutions, which are available on the market, were reviewed. The efficiency of the temperature measurements was tested by means of widely available thermal imaging cameras of a small format. It was determined that the requirements for the thermal imaging camera should be as follows: it has to be equipped with the SDK (Software Development Kit) which will enable its efficient operation from the level of a dedicated application. In addition, the temperature measurements must be accurate up to several metres of distance and have a resolution up to the tenth part of the degree. The camera for the project must be compact, mobile and operable by means of a separate device, for instance a mobile phone. For purposes of the project, two thermal imaging cameras, which are available on the market, were compared: Seek Thermal Pro (USB-C) and Flir One Pro LT (USB-C). The following features were analysed: – – – –

size and easiness of designing of inter-connected portable elements, available programming interfaces, measurement accuracy, quality of obtained images and processing efficiency.

The above analysis showed the supremacy of the Flir thermal imaging camera in many aspects important to the project. Due to this fact, further development of the project focused on this particular device. Also, available algorithms of the thermal imaging analysis were analysed. The efficiency of generally accessible algorithms

MONITORING HELMET ...

27

Fig. 1 An example of thermal images. Source own data

Fig. 2 Detection of a face with the mask in thermal imaging. Source own data

of the face detection was tested, also in the case of a partial covering of the face, for instance with a mask (Fig. 2), Within the framework of the current project, in order to process thermograms, the authors used a neural network which enables the detection of face. The Auto ML Vision Object Detection artificial neural network learned from thermal face images and identified the location of the face. This is an architecture developed by Google that allows for mobile and cloud-based object detection solutions. The MIT licence neural network [7] was trained on the basis of one set containing photos of faces taken in visible light (Cloud AutoML) and two sets including thermovisual images of faces (Tufts Face Database, FLIR ADAS Dataset). At the input, the model receives 3-channel images from the thermal imaging camera. These images were transformed into tensors of the following dimensions: 192 × 192. At the output, the following tensors are obtained from the network: coordinates for drawing a rectangle, class labels, certainty result for detected individual faces within a range of , the number of rectangles. Tensors are sorted according to the recognition certainty result, which means that the first index includes the detected face with the highest certainty result. Thanks to that fact, the verification of the certainty threshold may be limited to the first few results. In our implementation these were the first three results.

28

I. Chuchnowska et al.

The selected model may be implemented into the Android application on a mobile phone. The application processes images obtained through a small-format thermal imaging camera connected to the mobile phone. Prior to the implementation of the selected model into the Android application, the researchers carried out tests in Python in order to check the operation and accuracy of the facial recognition by the neural network. First of all, the network was tested on the basis of random photos of faces and figures (silhouettes) downloaded from the Internet. Next, images taken by a thermal imaging camera were tested. The last phase of the tests in Python was a video recording in thermovision. The model was able to process approximately 12 exposures per second with the settings of 4 themes in the i5-6400 2.7 GHz processor. The obtained results confirmed the utility of the network. The last stage following the face recognition is the search for the highest temperature within the outlined rectangle. If the temperature goes beyond the set threshold, the information is transmitted to the results presentation module. Within the framework of the project, it was decided to use three independent forms of measurements presentation enabling the user to choose the method best suited to their needs and limitations: – AR smart glasses, – transparent OLED display located on the visor of the portable device, – standard OLED display, from which the information will be reflected on a reflective foil located on the visor of the portable device. The requirements were formulated in relation to the equipment which could be used in the final solution: – AR smart glasses must have an open application programming interface (API) in order to be able to create own application for the operational system used by the AR glasses. In addition, they must have one of the wireless communication technologies to enable the transmission of data from the mobile phone to the glasses. – Device controlling the OLED display must be of small dimensions and charge a small amount of electric power to enable the power supply by a portable source of energy. In addition, the device must have wireless technology compatible with the technology used in mobile phone applications and an interface enabling the communication with the OLED display. – OLED display must be of a small size and be able to provide easy communication with its controlling device. For the needs of the project, two pairs of Augmented Reality smart glasses were compared: Epson Moverio and Vuzix Blade. The following features were analysed: operation of smart glasses, size and weight, image display, operational system, communication, creation of applications, and additional functions. On the grounds of the conducted analysis, the smart glasses Epson Moverio were chosen for purposes of this project. Within the framework of designing work, the use of two OLED displays was tested. The selected displays were as follows: a standard OLED display and a trans-

MONITORING HELMET ...

29

Fig. 3 Visualisation of possible versions of the Monitoring Helmet. Source own data

Fig. 4 Monitoring Helmet prototype. Source own data

parent SparkFun display. Both OLED displays are controlled by means of the I2C communication via Raspberry PI, which communicates with the mobile phone via Bluetooth. Tests were performed with a view to using helmets with thermal imaging camera. The equipment, methods of attachment and material solutions were analysed. Threedimensional models were prepared for purposes of the visual analysis of subsequent versions of the device as well as the creation of graphic and promotional campaign (Fig. 3). To ensure greater universality of the device, the authors designed a band which can be worn straight on the head or on any technical helmet. The band has a grip for mounting a thermal imaging camera and for possible attachment of a protective visor and OLED display. Additionally, the researchers designed a sachet making it possible to mount the designed casing for Raspberry PI as well as to hold the mobile phone and powerbank. During the designing works, the CAD models of load-bearing elements and assembly elements were developed. The above-mentioned models enabled the creation of the band prototype using 3D printing technology (Fig. 4).

30

I. Chuchnowska et al.

Fig. 5 Phases of identification process of individuals with fever by means of the Monitoring Helmet Source own data

4 Research Results The final effect of the project is a wearable device enabling the measurements of human body temperature at some distance using a thermal imaging camera which collaborates with a mobile phone in the AR technology (Fig. 4). The Monitoring Helmet is a portable device, which makes it possible to conduct measurements without using hands. The designed solution consists of mounting elements, a thermal imaging camera, a mobile phone and the AR display system. The elements are connected by means of wiring and constitute one system, whose components communicate with one another in order to achieve an intended goal. Following the start of a software programme, the camera begins to continuously collect thermal images, which partially overlap with the user’s field of vision. The gathered information is analysed and an attempt to take temperature of the person located closest to the user is undertaken. In the case of exceeding a certain threshold of temperature, the information on the detection of a dangerous person is sent to the user. The solution is controlled using a mobile application for the Android system. Subsequent phases of the process are presented in Fig. 5.

5 Conclusions This work presents a prototype of a mobile device (Monitoring Helmet) enabling the measurement of human body temperature at a distance. The application of the system of facial recognition in thermal images allowed an automatic identification of persons with the temperature exceeding a certain threshold. The implemented AR system makes it possible to automatically inform the user about the presence of individuals having their temperature above a certain pre-set temperature level, with a simultaneous minimum distraction of the user. The designed mounting system and load-bearing elements provide users with maximum physical freedom and universality of mounting. What is also important is that the set can be purchased at an affordable price and its components are widely accessible. The above-presented qualities of the device enable uncomplicated, quick and affordable implementation of

MONITORING HELMET ...

31

the Monitoring Helmet in the services provided by the personnel dealing with potentially ill people. The implementation of such equipment should provide increased protection in the working environment without hindering work of the device users. Acknowledgements This research was supported by the European Union from the European Social Fund in the framework of the project “Silesian University of Technology as a Center of Modern Education based on research and innovation” POWR.03.05.00-00-Z098/17.

References 1. Cholewka A, Drzazga Z, Siero´n A, Stanek A (2010) Thermovision diagnostics in chosen spine disease treated by whole body cryotheraphy. J Therm Anal Calorim 102(1):113–119 2. Kruszewski S (1971) Zastosowanie termografii i termowizji w medycynie. Pol Przegl Radiol Med Nukl 35(4):441–4 3. Praca zbiorowa pod redakcja˛ Madury H (2004) Pomiary termowizyjne w praktyce, Agenda wydawnicza Paku, Warszawa 4. Modrzejewska A, Parafiniuk M (2018) Zastosowanie termografii w medycynie - przegla˛d literatury. Pomeranian J Life Sci 64(3):29–32 5. Perpetuini D, Filippini Ch, Cardone D, Merla A (2021) An overview of thermal infrared imaging-based screenings during pandemic emergencies. Int J Environ Res Public Health 18(6):3286 6. Piccinini F, Martinelli G, Carbonaro A (2021) Reliability of body temperature measurements obtained with contactless infrared point thermometers commonly used during the COVID-19 pandemic. Sensors 21(11):3794 7. Braun M (2021) Thermal face project. https://github.com/maxbbraun/thermal-face. Accessed 20 May 2021

Electronic Stethoscope as a Tool Supporting Physician’s Work During the Covid-19 Pandemic Iwona Chuchnowska, Katarzyna Białas, Adam Gorol, Mateusz Gutkowski, Piotr Hałacz, Patryk Mielniczek, and Marta Pluta

Abstract The study presents a cordless stethoscope, which is a tool supporting physician’s work in time where direct contact with the patient is difficult or impossible. The study presents various types of available stethoscopes and their design. The stethoscopes presented in the article are important tools in physician’s work particularly now, at the time characterised by the limited possibility of patient examination. Difficulty in the examination of patients can also result from the necessity of using personal protective measures by medical personnel. The stethoscope helps monitor the health of patients who cannot leave the place of their stay. Keywords Stethoscope · Telephone/online medical consulting · Pandemic

1 Introduction The SARS-CoV-2 pandemic came as a surprise to people all over the world. Few people expected that the pandemic could change their everyday life, affecting both the global economy and lifestyle. Travelling is not allowed, contacts with other people should be reduced to a minimum. Companies capable of switching to online work have done so. Also physicians have been confronted with new challenges. Medical advice not requiring urgent and direct contact with the patient is provided by phone. The aforesaid situation has accelerated the development of technologies enabling remote or distant medical examination. The research involved the analysis and development of equipment, which could support physicians in their work taking into consideration risks resulting from direct contact with patients. The stethoscope is one of the primary “instruments”, by means of which the physician can diagnose the patient. Cordless stethoscopes are not an entire novelty, yet initial solutions were expensive. The authors of articles [1, 2, 5] analysed respiratory murmurs in healthy I. Chuchnowska (B) · M. Pluta Faculty of Biomedical Engineering, Silesian University of Technology, 41-800 Zabrze, Poland e-mail: [email protected] K. Białas · A. Gorol · M. Gutkowski · P. Hałacz · P. Mielniczek Faculty of Mechanical Engineering, Silesian University of Technology, 44-100 Gliwice, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_4

33

34

I. Chuchnowska et al.

patients and those with chronic obstructive pulmonary disease (COPD). The tests involved 32 patients suffering from COPD and 60 healthy individuals. The tests were performed using a stethoscope. The test results were subjected to analysis using a descriptive analytics software programme. The tests results were used to generate a linear amplitude spectrum, linear logarithmic amplitude spectrum and spectrogram of the patients. The stethoscope provided with a microphone was used to record acoustic signals of the heart. In article [2], the authors provided the description of sounds heard during proper auscultation and other frequently heard sounds as well as the meaning of each audible sound. The existing diagnostic solutions are not fully available because of their high price. The solution presented in this article could be used more commonly because of its low manufacturing cost.

2 Available Types of Stethoscopes The market offers a wide range of stethoscopes. One of the basic divisions is that into wired and wireless stethoscopes [3]. The primary components of the stethoscope include the head, a tube connecting the head with an eartube, eartubes (conducting the sound to each ear) and eartips (Fig. 1). The head consists of a drum and a diaphragm, being in a direct contact with the patient’s skin during examination. The drum, gently pressed against the thorax, transfers all sounds generated in its vicinity. Classical stethoscopes have a very similar structure. Differences usually occur within their heads. There are stethoscopes dedicated to adults and those to children (e.g. infants). Greater differences can be observed in electronic devices. The stethoscope, whose structure is the most similar to the classical one, is a Model 3200 electronic stethoscope (Fig. 2(a)). The most important difference concerns the head, which is fully electronic. The stethoscope also enables the recording and saving of up to 12 30-s long auscultation sounds. It can later play back the recorded sounds in 3 modes, i.e. a diaphragm, chestpiece or mixed mode. In addition, the stethoscope makes it possible to visualise recordings in the form of phonocardiogram. The stethoscope connects via Bluetooth. Another wireless stethoscope is named eKuore WIFI (Fig. 2(b)). This is a very accurate device which enables the comparison of patient’s examination results from different tests and makes it possible to reduce the number of medical checkups. Another example of an electronic stethoscope is Digital Stethoscope WISE (Fig. 3(a)). This stethoscope enables the auscultation of heart, lungs or intestines and features three sound analysis modes, i.e. a chestpiece, diaphragm and broad mode. The range of this wireless stethoscope is approximately five metres. The CloudSteth Smart Phone (electronic stethoscope) is one of the smallest stethoscopes on the market (Fig. 3(b)). It offers three modes of auscultation: heart, lungs and averaged (lungs + heart). The stethoscope features a system enabling sound volume control. In addition, the device switches off automatically if not in operation for 3 min. As a result of its unique design, the CloudSTETH electronic stethoscope is operated using the index finger only. The device also makes it possible to monitor the patient, record tests

Electronic Stethoscope as a Tool Supporting Physician’s Work ... Fig. 1 Classical stethoscope [6]

Fig. 2 Electronic stethoscope a model 3200 b eKuore WIFI

35

36

I. Chuchnowska et al.

Fig. 3 Electronic stethoscope a Digital WISE b CloudSteth Smart Phone

and save them for comparison at a later stage. Test results can be sent via the Internet to other users. There are also solutions based on artificial intelligence. An example of such a solution is a StethoMe wireless stethoscope. The device is adjusted to the current pandemic-related requirements concerning medical personnel, i.e. the need of using personal protective measures such as medical suits, masks or visors. Physicians using wireless stethoscopes can remotely monitor patients being in quarantine or in a risk group. After the pandemic, the StethoMe device will make it possible to monitor and provide assistance to patients without their leaving homes. All of the stethoscopes available on the market must satisfy basic normative requirements. Stethoscopes differ in terms of Bluetooth or Wi-Fi connection or the possibility of recording examination results for further analysis. Each physician chooses their stethoscope according to their likes and needs.

3 Description of Wireless Stethoscope Under Consideration The desire to ensure high availability and reduce costs of the stethoscope inspired an idea to print a stethoscope using a 3D printer (Fig. 5). The device will be controlled using a smartphone equipped with the Android system and a Bluetooth module. A special application will act as a receiver of signals transmitted by the stethoscope and enable its easy operation. During work on the interface, the designers decided to use a system where the main switch is located in the central part of the screen. The switch starts the reception of sound signals. The interface is presented in Fig. 4. Solutions providing wireless connection of the device included: Bluetooth, Internet and wi-fi standards. However, the use of Bluetooth is characterised by one disadvantage in terms of safety, i.e. the impossibility of providing connection over longer distances. To increase safety in the case of contagious diseases, the device could be connected using a Wi-Fi module. As a result, the physician would not need to contact the sick person directly, but use a mobile device. The device is easy to use and, after short

Electronic Stethoscope as a Tool Supporting Physician’s Work ...

37

Fig. 4 Application interface a before connection b connected

training, can be operated by the patient. The sounds captured by the stethoscope are sent via the Internet to the physician’s computer, where they can be subjected to analysis. One of the most important improvements is the possibility of saving sound recordings in the form of a file for archives or further analysis. To enhance the analysis of sounds sent by the wireless stethoscope, it is possible to use artificial intelligence. A related algorithm could analyse the frequency spectrum of the recorded data to detect murmurs, wheezes and other undesirable sounds and compare them with those stored in a previously created disease database. Supporting physicians with new technologies can improve the diagnostic process, detect potentially dangerous diseases and, in cases of larger groups of ill people, speed up necessary examinations. To adjust the application to current trends, the interface could be improved by adding an additional visual element, i.e. the so-called dark mode. The dark mode facilitates work in dark rooms and extends the battery operation time in devices provided with the OLED screens. Another improvement of the system could include the development of the application so that it could be used on devices supporting the iOS operating system. As a result, the device could be made available to a larger number of users. The electronic stethoscope in question is composed of the following elements: – housing and the microcontroller attachment made using the 3D printing process [4], – electronic components, – stethoscope head. The microcontroller controlling the stethoscope operation is fixed (by means of a movable fastening) in the central part of the housing. The fastening makes it possible

38

I. Chuchnowska et al.

Fig. 5 Wireless stethoscope printed using a 3D printer

Fig. 6 a Jack to connect headphones or a speaker b Switch (on/off)

to remove the microcontroller from the housing (e.g. when the microcontroller should be exchanged) or to modify soldered joints. The charging module is attached (using “hot” adhesive) to the upper wall of the housing directly above the microcontroller. The above-presented solution made it possible to provide the rear wall of the housing with a micro USB port (making it easily accessible). The rear wall is also provided with a switch and a jack bolted to the housing by means of a dedicated nut. The basket for one cell (18650) is attached to the bottom part of the housing. The sound amplifier is attached behind the microcontroller fastening. The stethoscope head has been attached to the microphone. In the aforesaid solution, the use of adhesive additionally seals the audio channel and, as a result improves the sound quality and reduces the outside noise. The entire head is attached to the rear wall of the housing. Since all of the electronic elements are soldered, no accidental disconnections should take place during the operation of the device. The joining of elements through soldering (instead of using the “gold pin” type of connections) enabled the obtainment of significantly smaller dimensions of the device. The rear and the front walls are bolted to the central part of the housing, which increases the durability of the device and makes it possible to access its internal elements (Fig. 5). The electronic stethoscope is simple to use. Persons who have not had contact with such a device in the past should not have any problems with using it. To perform a test it is necessary to do the following: 1) Plug in headphones or loudspeakers to the jack located in the rear part of the housing (Fig. 6(a)). 2) Switch on the stethoscope using the switch located next to the jack. The stethoscope is ready to auscultate the patient (Fig. 6(b)). 3) The front part of the stethoscope (where its head is located) should be put against the patient’s body. To eliminate noise caused by the patient’s hair, the device should be pressed firmly against the body.

Electronic Stethoscope as a Tool Supporting Physician’s Work ...

39

Fig. 7 a Lightning pictogram indicating the charger connection point b Pictogram of the screen indicating the micro USB port used for updating the device software

Fig. 8 Communication in the I 2 S protocol

After finishing the test, it is necessary to: 4) switch off the device, 5) unplug the headphones or the loudspeaker from the jack, and 6) disinfect the stethoscope head with antibacterial liquid. To charge the device, it is necessary to connect the charger using the micro USB port indicated with the pictogram of the lightning. The charging process is finished when the violet diode goes out (Fig. 7(a)). To update the software it is necessary to connect the device to a PC using a cable with a micro USB plug. The connection port is indicated with the pictogram of the PC screen (Fig. 7(b)). As regards the electronic part, two possible solutions were considered, i.e. the use of a Raspberry Pi microcomputer or an ESP32 microcontroller. Both platforms are provided with a dedicated converter enabling the execution of the task, yet the final solution was based on the ESP32 microcontroller due to its lower power consumption. The power supply system consists of an embedded 18650 li-ion battery having a capacity of 3500 mAh and a TP6045 controller, continuously monitoring the cell voltage and preventing both the discharge and the overcharge of the battery (potentially dangerous both for the user and the battery). The most important element of the stethoscope (critical for the quality of sound) is the microphone. The microphone used in the stethoscope, i.e. an SPH0645 model, uses the I 2 S protocol to communicate with the microcontroller (Fig. 8). The “SCK” line, i.e. Serial Clock (Bit Clock), is transmitted by the master device, i.e. a clock for transmitting and receiving bits.

40

I. Chuchnowska et al.

The frequency of the clock depends on the sound sampling frequency, the number of channels in use (left, right or both) and the number of bits used to record the sound (i.e. the level of compression). By means of the WS, i.e. Word Select line, the master device sets a low or high state (depending what channel is to be used for transmission by the slave device). Logical “0” is related to the left channel, whereas logical “1” is related to the right channel. The SD Serial Data line is used for the transfer of information in the form of bits compatible with the master device clock. The above-presented stethoscope was tested by a qualified physician. Due to fact that the sensitivity of the stethoscope is very high, the test results are very accurate. The jack makes it possible to connect the device to a loudspeaker or headphones, which, in turn, enables the adjustment of the sound volume. The greatest impediment to the quality of examination and sound purity could be patient’s excessive hair growth, which is responsible for the distortion of sound. The stethoscope under consideration can be used for auscultation of infants, children and adults. The use of a stethoscope is useful in cardiology, internally.

4 Summary Presently, when the pandemic makes direct contact with the physician difficult, direct appointments (at least some of them) must be replaced with telephone advice. Using the above-presented stethoscope, the physician can try to auscultate the patient via the telephone or computer. The patient possessing the stethoscope can individually auscultate themselves at home, simultaneously sending sounds to the physician. The proposed diagnostic solution could greatly facilitate physician’s work in present time. The stethoscope discussed in the article meets primary objectives related to it—the device can be used by physicians to carry out basic examinations without necessitating direct contact with patients. Further improvements to the device could include wireless data transmission or the automatic saving of the data in the computer storage.

References ˙ A, Zak ˙ M (2013) Cyfrowa analiza szmerów oddechowych 1. Grzegorczyk M, Dubaniewicz A, Zak u osób zdrowych i w przewlekłej obturacyjnej chorobie płuc. MONZ 19(2):89. (in Polish) 2. Maciuk M, Kuniszyk-Jó´zkowiak W, Ku´c L (2008) Analiza Fenomenów Osıuchowych. Sci Bull Chełm (1):95. (in Polish) 3. Malik B et al (2017) Design and development of an electronic stethoscope. In: Proceedings of the seventh international conference, pp 324–328 4. Marcisz D (2019) FDM - fused deposition modeling. https://mechatronikadlawszystkich.pl. (in Polish) 5. Wilk B (2007) Wirtualny stetoskop do badania tonów podstawowych serca. Pomiary automatyka kontrola (12):46. (in Polish) 6. https://medica91.com

The Epidemiological Analysis of the Atlantoaxial Joint Instability in Children and Youth with Down Syndrome Attending Special Facilities in Upper Silesia (Poland) Based on the Special Olympics Radiological Examination Iwona Doroniewicz, Paweł Linek, Andrzej Knapik, Adam Posłuszny, Aleksandra Masłowska, and Andrzej My´sliwiec Abstract Objectives Down syndrome is the most frequently occurring genotype aberrations in human with atlantoaxial articulation instability sometime combined with hypoplasia of the odontoid process. In this article, was aimed at estimating the size of the population of persons with DS in the Upper Silesia region and the selection of persons who have undergone atlantoaxial joint examination as part of Special Olympics’ preventative measures program. Material and Methods: Selected 31 towns in Poland a in which we sought special centers. Conducted the interview and analyze the necessary documentation. In all of the towns selected, the special facilities were located (61 total) and direct contact with the persons in charge of the facilities was obtained. Results: In total, 7105 people ranging in age from 7 to 21 years, Down syndrome occurred in 473 subjects, in 29 subjects conducted radiological examination and in 8 subjects was diagnosed atlantoaxial articulation instability. Conclusions: The obtained information should result in actions which would define the range of needs to conduct research concerning the problems for this population. The results achieved suggest that the number of examinations focused on the occurrence of atlantoaxial instability performed in persons with DS is insufficient. Keywords Down syndrome · Adapted physical activity · Special Olympic · Health prevention · Intellectual disability

I. Doroniewicz · P. Linek · A. Masłowska (B) · A. My´sliwiec Institute of Physiotherapy and Health Sciences, Academy of Physical Education in Katowice, Katowice, Poland e-mail: [email protected] A. Knapik Department of Health Care, Medical University of Silesia, Katowice, Poland A. Posłuszny The Joseph Tischner Special Schools Complex no. 10 in Jastrze˛bie-Zdrój, Jastrze˛bie-Zdrój, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_5

41

42

I. Doroniewicz et al.

1 Introduction Down syndrome (DS) is the most frequently occurring genotype aberration in humans, which leads to intellectual disability (ID) and which is the leading cause of specific birth defects and medical conditions [24]. It occurs in 1 out of 600–800 infants born alive. Approximately 20% of infants are born dead, while 60% of the cases comprise miscarriages in early stages of fetal life [12]. Among numerous disorders of the musculo-ligamentous-skeletal system in people with DS, the most characteristic include the functional and developmental anomaly within the region between the first (atlas, C1) and second (axis, C2) cervical vertebra, resulting in the underdevelopment of the axis dens. Most often, this anomaly is defined as atlantoaxial instability (AAI) or antlantoaxial subluxation (AAS). This issue affects about 6,8 to 27% of persons with DS [1, 2, 9, 11, 20, 25]. AAI was first mentioned in ancient times [20]. First screenings, however, were conducted in the 1960s by Spitzer et al. Interest in this topic increased in 1995, when the Committee on Sports Medicine of the American Academy of Pediatrics expressed its support towards the initiative of Special Olympics (SO), regarding the necessity of conducting radiological examinations of children with SD. The aim of this initiative was to prevent potential spinal cord injuries during sport and recreational activities [2, 11]. Currently, it cannot be unambiguously stated whether the occurrence of axis dense anomaly has any effect on the safety of persons during therapy, daily activities, recreational or sport activities. Nevertheless, the lack of these information cannot be deemed a restraining factor as far as preventative measures within this subject matter are concerned. Most of the research points to the occurrence of headaches and neck pain [1, 9, 14], although neurological consequences about the characteristics of palsy have also been noted in some studies [1, 18]. The first stage, presented in this article, was aimed at estimating the size of the population of persons with DS in the Upper Silesia region and the selection of persons who have undergone atlantoaxial joint examination as part of Special Olympics’ preventative measures program. The epidemiological data will enable us to indicate the percentage of persons with AAI in the studied population. The latter stage, dependent upon the previous ones, will be based on the radiological evaluation of persons with DS, who will be then provided with medical documentation regarding the condition of C1/C2 region. In this stage, it is planned for a head stabilizer to be prepared, in order to properly perform the X-ray scans. Currently, this task is usually performed by the parents, exposing them to the harmful effects of this procedure. In the final, upcoming and unfinished yet stage, the aim will be to evaluate, whether persons with DS differ functionally from healthy persons.

2 Methods The epidemiological stage of the study was based on the co-operation with SO Poland-Silesia, which is the only organization associating persons with DS in Poland

The Epidemiological Analysis of the Atlantoaxial Joint Instability ...

43

and which in the years 2000–2014 conducted X-ray examinations focusing on the occurrence of atlantoaxial joint instability. The whole project was approved by the Research Ethics Committee of The Jerzy Kukuczka Academy of Physical Education in Katowice (no. 7/2014 of 15 May 2014). 31 towns in the region of Silesia, in which special facilities attended by patients with ID (especially with DS) operate, were selected for the study. Some of these facilities partake in the Special Olympics program. Only the participants with valid sports medicine documentation and declaration of sports disciplines chosen for the upcoming training season were deemed ID patients with legitimate athlete status. In all of the towns selected, the special facilities were located (61 total) and direct contact with the persons in charge of the facilities was obtained. Through telephone conversations, specific meeting dates were arranged and persons responsible for providing information were chosen. In all of the cases, these were persons working as nurses or chairmans of Special Olympics sport clubs. These persons provided information in accordance with the participants’ medical documentation, as well as their Special Olympics sports medicine documentation handbooks. For specific details please see Table 1. All of the meetings took place on the premises of the facility, in a room in which conversation and analysis of medical documentation was possible. Following questions were asked: 1. How many patients are there in the facility (including men and women)? 2. How many patients with DS are there in the facility (including men and women)? 3. Is there a Special Olympics sport club operating at the facility? If not, do the patients participate in OS programs outside the facility? 4. How many of the patients hold SO athlete status? 5. How many patients with DS hold SO athlete status? 6. How many of the patients with DS received radiological examination focusing on axis dens hipoplasia? 7. In how many cases was atlantoaxial instability diagnosed?

3 Results 29 of the facilities (47% overall) co-operated with Special Olympics’ local agencies (Silesia). This co-operation consisted in the participation of persons with disability in regular training sessions and sport competitions. These persons were also the potential beneficiaries of other health organizations’ programs. From all of the facilities, a total of 966 (13,38%) persons with ID were reported to partake in Special Olympics programs on a regular basis (427 female participants—44,2%; 539 male participants—55,8%). Among these participants, 204 were persons with DS (47%; 83 female participants—40,7% and 121 male—59,3%). Amidst the persons with DS partaking in the Special Olympics program only in 29 cases altogether examinations were performed, which amounts to 14,2% of the population of persons with Special

44

I. Doroniewicz et al.

Table 1 Quantitative and percentile summary of the research material covered by the studies conducted in the Upper Silesia region (Poland) Town

Number of examined facilities

Total number of patients

Number Number of female of male patients patients

Number of persons with DS

Number of children with DS (%)

1

Be˛dzin

1

157

81

76

5

3.18

2

Bielsko-Biała

2

291

137

154

28

9.62

3

Bieru´n

1

340

186

154

2

0.59

4

Bytom

2

482

241

241

3

0.62

5

Chocznia

1

220

104

116

28

12.73

6

Cieszyn

1

193

105

88

11

5.70

7

Czelad´z

1

60

26

34

8

13.33

8

Cze˛stochowa

5

292

161

131

34

11.64

9

Da˛browa Górnicza

5

239

95

144

22

9.21

10

Gliwice

1

212

80

132

10

4.72

11

Jastrze˛bie-Zdrój

2

209

68

141

16

7.66

12

Jaworzno

2

253

115

138

11

4.35

13

Katowice

4

349

141

208

40

11.46

14

Lubliniec

1

214

100

114

9

4.21

15

Mikołów

2

138

67

71

13

9.42

16

Milówka

1

28

10

18

6

21.43

17

Mysłowice

1

138

56

82

8

5.80

18

Myszków

2

49

27

22

2

4.08

19

1

196

88

108

18

9.18

20

Racibórz ´ ˛ska Ruda Sla

3

291

124

167

15

5.15

21

Rybnik

3

291

117

174

35

12.03

22

Skoczów

1

71

36

35

7

9.86

23

6

886

426

460

51

5.76

24

Sosnowiec ´ ˛tochłowice Swie

2

179

75

104

3

1.68

25

Tarnowskie Góry

1

178

80

98

10

5.62

26

Tychy

2

253

99

154

21

8.30

27

Ustro´n

1

90

39

51

7

7.78

28

´ ˛ski Wodzisław Sla

1

245

92

153

28

11.43

29

Zabrze

3

271

104

167

10

3.69

30

Zawiercie ˙ Zywiec

1

132

56

76

5

3.79

1

158

88

70

7

4.43

61

7105

3224

3881

473

6.66

31

Total

The Epidemiological Analysis of the Atlantoaxial Joint Instability ...

45

Table 2 Quantitative and percentile summary of achieved results Test parameter n Percentage % Examined facilities 61 Facilities realizing the SO 29 program Persons with SO athlete status 966 Persons with DS participating 204 in SO programs Persons with DS and AAI 29 Persons diagnosed with AAI 8

100 47.54 100 47 14.2 27.59

Olympics athlete status and only 6,61% of the general population of people with DS attending the facilities in the Upper Silesia region, who were selected for the study. Atlantoaxial instability was diagnosed by a physician radiologist in 8 cases, which amounts to 27,59% of the radiologically examined population. For specific details please see Table 2. The results achieved enabled us to perform a forecasting necessary to conduct comprehensive population studies of persons with DS, focusing on the occurrence of antlantoaxial instability in the population of persons with DS living in the Upper Silesia region. A group of 269 persons with DS (56,8%) does not participate in Special Olympics and other health programs. It is necessary to conduct radiological examinations among SO athletes in 180 cases (88%), and as far as the entire studied population of persons with DS is concerned—in 449 cases. Assuming the percentage indicator of atlantoaxial instability occurrence to be 27,59% (based on the results achieved), it can be indicated that among the unexamined population of persons with DS, AAI may occur in about 120 of the cases.

4 Discussion Performing atlantoaxial examination of persons with DS is a complex task. Apart from the financial capabilities of Regional Agencies of Special Olympics in Poland, also proper education of trainers, parents and radiological laboratory staff is essential. It is not infrequent, that persons with DS experience medical phobia. They find it hard to remain calm during the procedures, not to talk about the necessity of placing one’s head in a fixed position, with mouth and eyes open. These difficulties may lead to medical assessments being subject to the risk of error. The SO regulations clearly state which sports disciplines are contraindicated for persons with AAI and unexamined persons. These are: butterfly (swimming style), starting dive, pentathlon, high jump, horseback riding, barbell squat (biathlon, triathlon), artistic gymnastics, football, alpine skiing, neck warm-up exercises [27]. So wide a list of contraindicated sport

46

I. Doroniewicz et al.

activities, including unitary activities, which are a part of physiotherapy protocols, physical education and touristic forms of recreation, accentuates the need for further study of this problem. In the radiological examination, this anomaly is indicated by the extension of the distance between the rear surface of the anterior arcus of C1 and the anterior surface of the C2 dens. The measurement of this distance indicates the characterization used for describing the anomaly. This distance should be measured in forward flexion of the head, in a neutral position and in straight posture, and it should amount to less than 4 mm [5], 4,5 mm [16], up to 5 mm [21]. Additionally, Burke et al. suggest a value of less than 3 mm for persons under the age of 15 [4]. As far as the world literature on this topic is concerned, it needs to be noted that the methodological problems regarding the research on AAI are not limited exclusively to the contradistinction between “instability” and “subluxation”. The researchers also criticize the existing methods of proper X-ray examination, which need to include maximal range of motion in flexion and in extension of the examined joint. This requires proper positioning of the person examined, stabilization of other parts of the spine and co-operation with the patient. Taking into account various cognitive disorders and intellectual disability of different stages, often co-existing with DS, which affect the communication with such persons, the imaging examinations are subject to errors foreclosing the ability to properly assess the anomaly [3, 7]. A result indicating instability can be affected by, for example, upper respiratory tract infection, which can lead to loosening of the transverse ligament of the first cervical vertebra, muscle tension resulting from emotional state and mood, post-tonsillectomy state or juvenile idiopathic arthritis [8]. Also, weak correlation between AAI results and the patient’s neurological symptoms, indicating pressure on the spinal cord, are said to be another drawback of the radiological examination [6, 10]. Low dependency is additionally caused by the fact, that weariness, gait disorders, limited cervical spine mobility, torticollis, loss of coordination, dysesthesia, spasticity, overreactivity, muscle clonuses, excessive myotatic contraction, muscle weakness, decreased sphincter and bladder control are actual and serious issues, affecting 1–5% of the population of persons with DS [15, 22, 26]. The only cause of AAI is not the loosening of the C1 vertebra transverse ligament but—as suggested by the researchers—abnormality in the axis dens structure [17, 19]. In the aforementioned research, dysplastic dens of the axis is described by the shortening of the dens below the value of two standard deviations of the average height of the dens in healthy persons. The height of the dens was measured using the McManners’ method (1983), which consists in measuring the distance between the apex dentis and the upper surface of the C2 vertebra. Elliott et al. [8] additionally measured the so-called C2 axis dens deficit, that is the distance between the apex dentis and the upper edge of the C1 arcus anterior. Both these values were comparable to the values reached in the group of healthy persons of the same age. As indicated by the research, while the comparison of both distances measured suggested a tendency toward hypoplasia in persons with DS, the difference was statistically significant only in the adult group. Up to the age of 9 years the values of distances measured were

The Epidemiological Analysis of the Atlantoaxial Joint Instability ...

47

similar. These conclusions remain in line with what was suggested by Selby et al. [23], who stated that the age of 8–9 years is the boundary value for the development of the C2 axis dense. Long-term research on AAI indicated, that the extent of this anomaly tends to reduce with age [5, 13]. Therefore, the results of the aforementioned studies seem to suggest, that in persons with DS who suffer from C2 axis dens hypoplasia and AAI simultaneously, a compensative mechanism, which increases the coherence of this segment of the spine and prevents the onset of neurological symptoms, is developed. This data should lead to the undertaking of activities aimed at estimating the need of conducting research on the issues of this population and obtaining adequate means in order to do so. Such endeavors shall definitely contribute to the improvement of social functioning of persons with DS, and even more importantly, their quality of life. Lastly, resolving the doubts of physical education teachers, adapted physical activity instructors and trainers, regarding the choice of safe and adequate sports disciplines, would be another positive effect.

5 Conclusions 1. The conducted research indicates the occurrence of atlantoaxial instability in persons with DS living in the Upper Silesia region (Poland). 2. The results achieved suggest that the number of examinations focused on the occurrence of atlantoaxial instability performed in persons with DS is insufficient. 3. The results achieved enable us to estimate the expenditures needed for the realization of further stages of the “More stable, more physically fit” program.

References 1. Ali FE, Al-Bustan MA, Al-Busairi WA, Al-Mulla FA, Esbaita EY (2006) Cervical spine abnormalities associated with Down syndrome. Int Orthop 30:284–289. https://doi.org/10.1007/ s00264-005-0070-y 2. American Academy of Pediatrics Committee on Sports Medicine and Fitness (1995) Atlantoaxial instability in Down syndrome: subject review. Pediatrics 96(1):151–154 3. Bajaj M, Jangid H, Vats ML (2010) Congenital absence of the dens. Indian J Radiol Imagin. 20(2):109–111. https://doi.org/10.4103//097-3026.63050 4. Burke SW, French HG, Roberts JM et al (1985) Chronic atlanto-axial instability in Down syndrome. J Bone Joint Surg Am 67:1356–1360 5. Cremers MJG, Ramos L, Bol E, van Gijn J (1993) Radiological assessment of the atlantoaxial distance in Down’s syndrome. Arch Dis Child 69:347–350. http://dx.doi.org/10.1136/adc.69. 3.347 6. Davidson RG (1988) Atlantoaxial instability in individuals with Down syndrome: a fresh look at the evidence. Pediatrics 81:857–865 7. Elliott C (1988) The odontoid process in children: is it hypoplastic? Clin Radiol 39:391–393. https://doi.org/10.1016/s0009-9260(88)80278-2

48

I. Doroniewicz et al.

8. Elliott S, Morton RE, Whitelaw RAJ (1988) Atlantoaxial instability and abnormalities of the odontoid in Down’s syndrome. Arch Dis Child 63:1484–1489. https://doi.org/10.1136/adc.63. 12.1484 9. Ferguson RL, Putney ME, Allen BL Jr (1997) Comparison of neurologic deficits with atlantodens intervals in patients with Down syndrome. J Spinal Disord 10:246–252 10. Jagjivan B, Spencer PA, Hosking G (1988) Radiological screening for atlanto-axial instability in Down’s syndrome. Clin Radiol 39:661–663. https://doi.org/10.1016/s0009-9260(88)800886 11. Kattan H, McDonald P (1996) Atlanto-occipital and atlanto-axial instability in children with Down syndrome. Ann Saudi Med 16:56–59. https://doi.org/10.5144/0256-4947.1996.56 12. Korenberg JR, Chen XN, Schipper R et al (1994) Down syndrome phenotypes: the consequences of chromosomal imbalance. Proc Natl Acad Sci USA 91:4997–5001. https://doi.org/ 10.1073/pnas.91.11.4997 13. McManners T (1983) Odontoid hypoplasia. Br J Radiol 56(672):907–910 14. Miller JD, Capusten BM, Lampard R (1986) Changes at the base of skull and cervical spine in Down syndrome. Can Assoc Radiol J 37:85–89 15. Msall ME, Reese ME, DiGaudio K et al (1990) Symptomatic atlantoaxial instability associated with medical and rehabilitative procedures in children with Down syndrome. Pediatrics 85(3):447–449 16. Onerci M, Ogretmenogolu O (1997) Atlantoaxial subluxation after tonsillectomy and adenoidectomy. Otolaryngol Head Neck Surg 116(2):271–273. https://doi.org/10.1016/s01945998(97)70341-3 17. Pueschel SM, Herndon JH, Gelch MM, Senft KE, Scola FH, Goldberg MJ (1984) Symptomatic atlantoaxial subluxation in persons with Down syndrome. J Pediatr Orthop 4:682–688. https:// doi.org/10.1097/01241398-198411000-00005 18. Pueschel SM, Scola FH (1987) Atlantoaxial instability in individuals with down syndrome: epidemiologic, radiographic and clinical studies. Pediatrics 4:555–556 19. Pueschel SM, Scola FH, Perry CD, Pezzullo JC (1981) Atlanto-axial instability in children with Down syndrome. Pediatr Radiol 10:129–132 20. Pueschel SM, Scola FH, Pezzullo JC (1992) A longitudinal study of atlanto-dens relationships in asymptomatic individuals with Down syndrome. Pediatrics 89:1194–1198 21. Pueschel SM, Scola FH, Tupper TB, Pezzullo JC (1990) Skeletal anomalies of the upper cervical spine in children with Down syndrome. J Pediatr Orthop 10:607–611. https://doi.org/ 10.1097/01241398-199009000-00007 22. Roy M, Baxter M, Roy A (1990) Atlantoaxial instability in Down syndrome - guidelines for screening and detection. J R Soc Med 83:433–435 23. Selby KA, Newton RW, Gupta S, Hunt L (1991) Clinical predictors and radiological reliability in atlantoaxial subluxation in Down’s syndrome. Arch Dis Child 66(7):776–778. https://doi. org/10.1136/adc.66.7.876 24. Sherman SL, Allen EG, Bean LH, Freeman SB (2007) Epidemiology of Down syndrome. Ment Retard Dev Disabil Res Rev 13(3):221–227. https://doi.org/10.1002/mrdd.20157 25. Spitzer R, Rabinowich JY, Wybar KC (1961) A study of abnormalities of the skull, teeth and lenses in mongolism. J Can Med Assoc 84:567–572 26. Tredwell SJ, Newman DE, Lockitch G (1990) Instability of the upper cervical spine in Down syndrome. J Pediatr Orthop 10:602–606. https://doi.org/10.1097/01241398-199009000-00006 27. Special Olympics [Internet] (2014). General rules of special olympics. http://www. specialolympics.org. Accessed 14 Oct 2014

Trends in Development of Balance Dysfunctionalities Rehabilitation Equipment Using Virtual Reality—A Literature Review Grzegorz Gruszka, Piotr Wodarski, Marek Ples, Marta Chmura, Andrzej Bieniek, and Jacek Jurkoj´c Abstract Introduction Rehabilitation of balance maintaining ability is a longlasting process, requiring strict cooperation between physiotherapist and the patient. Rehabilitation process is usually monotonous, requires a lot of dedication and is just not attractive for patients, while allowing physiotherapists to assess the progression of the process qualitatively only. Used in traditional physical rehabilitation methods of obtaining quantitative data can be insufficient to objectively evaluate if any improvement has been achieved or changes in rehabilitation program must be made. Purpose Virtual reality (VR) technology is becoming a trending solution for said inconveniences and problems, allowing patients to become immersed in an entertaining, unreal world while still performing rehabilitative tasks and receiving immediate feedback on progress that is being made. Authors aim to establish trends in development of VR using systems used for enhancing rehabilitation of balance maintaining ability, which will make it easier for physiotherapists to pick suitable equipment, and for companies creating rehabilitation equipment to create devices with greater potential of being a successful inventions. Material and Methods Authors reviewed 110 publications found in databases of Pubmed, Researchgate, Mendeley (prior to version 1.19.8) and Google Scholar. After applying inclusion and exclusion criteria 23 publications were put into further considerations. Results and Conclusions Reviewed systems were classified into 3 groups and analyzed. 15 publications were created using more simple, market available systems, 5 were using custom-made devices and 3 made use of highly advanced and not-easily accessible systems. Trends in development of VR using rehabilitation devices have been established in favour of more simple, but easily accessible systems. Keywords Diagnostic · Human gait · Virtual reality · Rehabilitation · Balance

G. Gruszka (B) · P. Wodarski · M. Ples · M. Chmura · A. Bieniek · J. Jurkoj´c Faculty of Biomedical Engineering, Silesian University of Technology, 40 Roosevelta Street, 41-800 Zabrze, Poland e-mail: [email protected] URL: https://www.polsl.pl/rib3/en/ © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_6

49

50

G. Gruszka et al.

1 Introduction Rehabilitation is a process which often uses various devices, purpose of which is to make cooperation between physiotherapists and patients easier, and cause entire therapy to be more, if not at all, effective. Process of regaining physical capabilities or alleviating dysfunctionalities and impairments takes very long time, is monotonous and simply not attractive for patients, lowering their motivation and will to complete entire therapy [1]. Each patient has to be supervised by at least one person monitoring the rehabilitation process, and the amount of supervisors can be greater, depending on severity of dysfunction or impairment. This leads to increase in costs of therapy and delays the start of rehabilitation. Furthermore, postponing the start of rehabilitation can result in lesser efficiency of recovery process [2], which in turn can further extend the required rehabilitation time, thus lowering the chance of patient completing the entire rehabilitation process even more. A significant drawback present in conventional rehabilitation is relatively small accessibility to quantitative data and feedback—process of performing the exercises is often evaluated only qualitatively by either the physiotherapists during assisted exercises, or the patient when he/she exercises at home. The latter is even worse, since patients usually do not have experience to draw conclusion whether they are performing the exercises correctly. It is also difficult to perform self-evaluation of performed exercises, especially when there is no other person to monitor if exercises are performed properly. Quantitative data is usually obtained only during check-ups of rehabilitation progression, through use of devices such as AMTI or ZEBRIS force platforms, which allow checking if improvement has been achieved by performing measurements of centre of pressure (COP) and force distribution, but do not give visual, appealing feedback during exercises. This affects patients as well, since they are unable to directly see their improvements and might be exercising improperly due to lack of feedback. Said platforms serve research or diagnostic purposes and therefore do not contain any entertaining aspects. Immediate access to quantitative data obtained during the exercises can help achieve progress more efficiently, shortening the time necessary for completing entirety of rehabilitation process. Quantitative data allows physiotherapists to objectively evaluate if the exercises patients perform are giving positive results and promise as efficient recovery as possible, or if changes in the rehabilitation program should be made in order to increase its effectiveness. For patients, such data combined with real time feedback lets them immediately see the progress they are making, allows introduction of entertaining aspects and therefore increases their will to complete the rehabilitation in its entirety [3, 4]. In past two decades Virtual reality (VR)/Augumented Reality (AR)/“exergaming” has been becoming more and more recognizable trend in rehabilitation. In 2004 it was already established that VR is a promising tool for enhancing rehabilitation process [5], therefore it can be stated that introduction of VR to conventional rehabilitation is either beneficial or neutral both to patients and physiotherapists, making it unreasonable not to use it during physical therapy. It has been further confirmed bt Cano Porras et al. in [3]. Introduction of VR has potential to improve scores on

Trends in Development ...

51

evaluation tests, increases patient’s engagement, motivation and possible intensity of training [6–10]. Lately more and more conventional systems, such as treadmills or rehabilitation robots are getting coupled with VR or AR in variety of variants—from CRT and regular PC monitors, through widescreen TVs and projectors to Google Glass™ [11] and head mounted displays (HMD) [12]. What is worth noting is relative easiness of coupling VR with rehabilitative devices such as treadmills, robots or balance boards. In most cases no changes to already available devices have to be made in order for it to cooperate with unreal world. Virtual reality also allows high personalization of entire rehabilitation process. It is possible to introduce appealing scenarios, and with properly made software, adjustments to exercises can be made “on the spot”. This paper aims to review and establish trends in development of balance maintaining ability physical rehabilitation VR using equipment based on what kinds of devices are currently used for performing research or rehabilitating human balance maintaining ability. Rapid development of physical rehabilitation enhancing systems leads to increase in variety of available devices. Nowadays research can be performed on highly advanced, expensive and accessible only in certain places in the world measuring systems, as well as on mobile and widely available devices used primarily for entertainment purposes. Establishing these trends will allow physiotherapists to select suitable to their needs systems, gain knowledge on what kinds of devices and how are they used in order to increase efficiency of rehabilitation process, and for companies manufacturing rehabilitation equipment or working in field of physical rehabilitation this review will serve as a guideline for what types of devices are demanded. For researchers this paper aims for giving insight on what kinds of devices can be used for purposes of performing scientific measurements with satisfying accuracy and sufficient quality, which can later serve for creating new physical rehabilitation concepts, programs or exercises.

2 Methodology 2.1 Search Strategy Reviewed publication bases include Pubmed, Researchgate, Mendeley (prior to version 1.19.8) and Google Scholar. Searched strings were: “VR rehabilitation”, “VR ‘AND’ rehabilitation”, “Virtual reality rehabilitation”, “virtual reality ‘AND’ rehabilitation”, “Exergaming”, “VR Balance”, “Virtual Reality balance rehabilitation”, “Virtual Reality Balance”.

52

G. Gruszka et al.

2.2 Search Strategy Papers which were included in our review had to: – Be not older than from 2010. – Research use of VR/AR or “exergaming” in rehabilitation of balance maintaining ability of human beings. – Use VR/AR equipment independently or regular rehabilitation devices fitted with VR/AR equipment for purposes of rehabilitation or diagnosing human balance maintaining ability. – No disability or impairment was focused on. Duration of performed research, amount of participants were also not focused on (did not serve as a criteria for inegibility). – If custom made devices were used their construction had to be described as well as way in which the VR/AR equipment was used in them. Papers were excluded if: – The principle of working of custom made systems was not described – There was no indication of using VR/AR equipment – No statement about differences between pre- and post-rehabilitation results was present – Methods of evaluating the differences between pre- and post-rehabilitation were not presented – Custom made exercise applications were not described – Custom made measuring applications for obtaining data for later analysis were not described

2.3 Review Process Authors found 110 articles after searching mentioned publication bases. All of found articles were first briefly revised by authors. All duplicates were removed, so were responses of papers’ authors to questions regarding their works and reports confirming the quality of their work. Appendixes to works which were not found in our search results were also discarded. Due to before mentioned circumstances, 19 publications were not put into any further consideration. After initial discarding of unsuitable publications, 91 papers were thoroughly screened and underwent assessing through inclusion and exclusion criteria. No conflict between authors arose about inclusion or exclusion of any of the remaining articles. Screening of remaining articles left 23 publications which were qualified to being included into further review.

Trends in Development ...

53

Fig. 1 PRISMA flowchart representing the revision process [13]

3 Results After identifying, screening through and applying all the inclusion criteria 23 publications were found to meet all the requirements. Authors, types of measuring systems, their components, methods of evaluation of effects of physical rehabilitation using VR/AR and final conclusions regarding the effects of introducing VR to rehabilitation were put in Table 1 to present what are the most important in authors opinions aspects of performed review. All systems used in reviewed publications were classified by authors to one of three groups: market-available, custom-made and advanced. Criteria for classification were as follows (Fig. 1):

54

G. Gruszka et al.

Table 1 Characteristics of reviewed systems STS—Sit to Stand transfer; BBS—Berg’s Balance Scale; TUGT/TUG—Timed up and go test; MMAS—Modified motor assessment scale; FIM— Functional Independence measure; FES-I; Short-form falls efficiacy scale-international; FAB; Fullerton Advanced Balance Scale; FRT Functional Reach Test; MBI Modified Barthel Index; 10MWT 10 m-walk test; SBT Standing Balance Test; FGA—Functional Gait Assessment; PASS— Postural Assessment Scale; DGI—Dynamic Gait Index Reviewed paper Measuring Evaluation Influence of Authors’ system/-s methods of introducing VR classification of rehabilitation on results of used system/-s results participants [7]

Microsoft Kinect™

– BBS –TUGT –FGA

[14]

Set of CAREN systems: –Dome –Base –V-Gait –C-Mill Coupled with motion capture system (Vicon, Oxford, UK) BTS Nirvana VR Interactive System

–Berg’s Balance Scale (BBS) –miniBESTes

DTVRBT— custom system consisting of a PC and a Kinect controller

–Calculations performed in MATLAB environment

[15]

[16]

–BBS –TUGT –FGA –Falls Efficacy Scale International

Beneficial in Market-Available favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control groupy Beneficial in Advanced favour of VR enhanced therapy— registered better results in balance evaluation after therapy

Neutral—no Advanced significant differences between control group (no-VR training) and experimental group were found Beneficial in Custom-made favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group (continued)

Trends in Development ... Table 1 (continued) Reviewed paper Measuring system/-s

55

Evaluation methods of rehabilitation results

Influence of introducing VR on results of participants

Beneficial in Market-Available favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group Beneficial in Market-Available favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group Neutral—no Advanced significant differences between control group (no-VR training) and experimental group were found

[17]

Nintendo Wii console, 30 TV, Wii Sports game

–PASS –MMAS –FIM

[18]

42 TV, Wii Balance Board

–BBS –TUGT

[19]

BRU (balance rehabilitation unit) + “virtual reality glasses”

[20]

–Through use of BRU: Standing with eyes open, standing with eyes closed, standing on a foam with eyes closed, saccadic task Wii Fit game with –BBS WIi Balance –Bubble Test Board (one of Wii games)

Authors’ classification of used system/-s

Neutral as Market-Available standalone system due to small registered changes, beneficial in combination with physical exercises as entertainment feature (continued)

56

G. Gruszka et al.

Table 1 (continued) Reviewed paper Measuring system/-s

Evaluation methods of rehabilitation results

[21]

Wii Fit game with –BBS WIi Balance –TUG Board –FAB –FRT

[22]

Wii Fit game with –Romberg’s test WIi Balance on Bio-Rescue Board

[23]

Wii Fit game with –BBS WIi Balance –Tinetti Gait and Board Balance Assesment

[24]

Wii Fit game with –TUG WIi Balance –FES-I Board –30 s repeated chair stand test

Influence of introducing VR on results of participants

Authors’ classification of used system/-s

Beneficial in favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group Beneficial in favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group Neutral—no significant differences after VR intervention; VR designated as appealing enhancement introducing entertainment to rehab. Neutral—no changes in balance maintaining ability recorded. Beneficial as adherenceenhancing device. Recorded increase in isometric muscle strength

Market-Available

Market-Available

Market-Available

Market-Available

(continued)

Trends in Development ... Table 1 (continued) Reviewed paper Measuring system/-s

57

Evaluation methods of rehabilitation results

Influence of introducing VR on results of participants

Authors’ classification of used system/-s

[25]

PC with a 22 touchscreen, connected with a wireless stabilographic platform

[26]

Beneficial in Market-Available favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group Classic treadmill –BBS Beneficial in Custom-made coupled with a –TUG favour of VR PC. Displayed –Postural Sway enhanced were recordings (with the pressure therapy— of real world with platform) registered better use of a projector results in balance evaluation after therapy in comparison with control group Classic treadmill –BBS Beneficial in Custom-made coupled with a –TUG favour of VR PC. Displayed enhanced were recordings therapy— of real world with registered better use of a projector results in balance evaluation after therapy in comparison with control group

[27]

[28]

–BBS –DGI –TUG –Parkinson’s Disease Questionnaire –Unified Parkinson’s Disease Rating Scale Wii Fit game with –BBS WIi Balance –10MWT Board

Neutral—no Custom-made significant differences between control group (no-VR training) and experimental group were found

(continued)

58

G. Gruszka et al.

Table 1 (continued) Reviewed paper Measuring system/-s

Evaluation methods of rehabilitation results

Influence of introducing VR on results of participants

Authors’ classification of used system/-s

–Measurements on pressure platform –STS

Beneficial in favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group Beneficial in favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group Beneficial in favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group Neutral—no significant differences between control group (no-VR training) and experimental group were found; VR group described experience as “more pleasant”

Custom-made

[29]

Classic treadmill coupled with a PC, custom made games displayed on a LCD screen, two switches on sides of treadmill

[30]

Wii balance –BBS board-based –Anterior system (eBaViR) Reaches Test

[31]

Wii Fit Balance –FRT Board; Microsoft –BBS Kinect™ –TUG –Measurements of COP –MBI

[32]

Microsoft Kinect™

–BBS –FRT –TUG

Market-Available

Market-Available

Market-Available

(continued)

Trends in Development ... Table 1 (continued) Reviewed paper Measuring system/-s

59

Evaluation methods of rehabilitation results

Influence of introducing VR on results of participants

Beneficial in Market-Available favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group Neutral —no Market-Available significant differences between control group (no-VR training) and experimental group were found; VR training proved more cost-efficient Beneficial in Market-Available favour of VR enhanced therapy— registered better results in balance evaluation after therapy in comparison with control group

[33]

Wii Fit Balance Board

–BBS –TUG

[34]

TV screen, PC, Kinect™102 (Microsoft®, WA), 42 LCD screen

–BBS

[35]

Nintendo Wii –Balancia Sports Resort + software + Wii canoe-like Balance Board apparatus, created by fixing a chair to a springboard

Authors’ classification of used system/-s

– Market-available: VR devices used in research are available to be bought in stores and do not require any interference from researchers. Third-party software used with available in stores equipment was not disqualifying from being classified as Market-available. Such systems can be used at homes by patients themselves, are relatively easy to obtain, their main intended purpose is entertainment, and are usually relatively cheap. Such systems are Nintendo Wii Balance Board and Microsoft Kinect. – Custom-made: VR devices used in research are made from external parts, assembled for purposes of performing the research. Devices from this group are not widely available for purchase.

60

G. Gruszka et al.

– Advanced: Complex systems accessible only in few places in the world, not widely available and possible to purchase only from their producers or representatives. Such devices are made purely for purposes of performing research on biomechanics while utilizing possibilities of VR. As can be seen in Table 1, systems used for physical rehabilitation and diagnosis of human balance maintaining ability vary from taking a lot of space, using highquality, advanced components and capable of mimicking complex scenarios, through custom-made devices made by enhancing already existing devices such as treadmills with HMD or regular screens, to relatively simple, easily accessible and available in RTV stores, main purpose of which is entertainment. Said variety confirms complexity of restoring balance maintaining process and diversity of possible approaches to the problem. Out of 23 reviewed publications 15 state that introducing VR to rehabilitation programs or replacing regular rehabilitation with VR exercises was clearly beneficial for the patients and gave better results in comparison with regular, non-VR physiotherapy. Remaining 8 publications stated that introduction of VR equipment was neutral and did not differ significantly in results from progress achieved through regular physiotherapy. It must be noted although, that even when no significant differences were apparent between the VR and Non-VR groups enhancement of regular physiotherapy is still recommended because of the appealing scenarios and entertaining introduced with VR. Out of 15 systems classified as “Market-available” 10 were considered as beneficial due to increasing evaluation scores of rehabilitation progress. Out of 5 systems classified as “custom-made” 4 declared clear benefit of introducing VR to rehabilitation. From 3 publications making use of systems classified as “advanced” only one reported clear benefit favouring use of VR.

4 Discussion Review of systems used for rehabilitation of balance maintaining capability in human beings allows selection of trends appearing in physical rehabilitation environment. Most of the reviewed articles were using relatively simple, widely available devices such as Nintendo Wii Balance Board or Microsoft Kinect (fifteen out of twentythree reviewed works). Devices created purposefully for rehabilitation or diagnosis of balance maintaining ability appear to be used less frequently—only three papers were found in span of eleven years. Reasons behind such outcome can vary, but it is noticeable that systems used more frequently for purposes of commencing research on balance maintaining ability are easily accessible (Wii Balance Board and Kinect controllers are still available to buy in stores), are made not solely for rehabilitation, and their intended purpose is to serve as entertainment devices. They do not take a lot of space and allow quick and relatively easy customisation/adaptation to research requirements, and even can be used for rehabilitation without any interference. Expensive systems made solely for research purposes, or devices consisting of many parts and taking a lot of space are not used often. They are not widely available

Trends in Development ...

61

for patients nor for therapists, making use of smaller, easier to access systems a more reasonable choice. Another aspect worth of noting is that all the “market-available” devices allow patients to train comfortably at homes without necessity of regular appearance on rehabilitation sessions in according facilities, which takes time and can be close to impossible if patient struggles with commuting or does not have means of getting to such facility. What is also important while comparing the easily accessible, small and cheap systems to big, stationary and made purely for research or rehabilitation of balance maintaining systems is the results reported by authors performing the research. Review of all of listed works allows us to state that none of the reviewed publications reported performance of Wii Balance Board or Microsoft Kinect to be insufficient to receive valuable and useful data or perform the measurements, nor influenced the rehabilitation process in any negative way. All of reviewed works reported beneficial or neutral influence, with a note of VR still being a valuable addition due to its entertaining aspects increasing patients engagement and will to finish entire process while not affecting the rehabilitation in any negative way. It seems that “advanced” systems are less desirable or are on decline, while mobile, small and easily accessible systems are becoming the future of balance maintaining rehabilitation and diagnosis, which is proved in [36, 37]. It cannot be omitted although, that due to low accessibility of the “advanced” systems fewer research works resulting in publications are done. Their great advantages over the simpler systems are their security features and possibility of making research on scenarios which would otherwise be difficult to replicate in either real life or through use of simpler systems. What follows is possibility of exploring reactions to new kinds of stimuli conflicts, or discoveries of new rehabilitative or diagnostic techniques. Quality of components and measurements is also much better than what is obtainable with cheaper systems, allowing researchers and therapists to monitor changes and occurrences that would otherwise remain unnoticed. For example, research of Anticipatory Postural Adjustment (APA) is easier to notice if not made possible to begin with by using the more expensive and advanced systems due to way higher quality and greater possibilities of components used in production of the “advanced” systems. Wii Balance Board allows performing measurements 40 Hz sampling frequency, which becomes serious issue when trying to measure and analyze short lasting phenomena as APA, which lasts from around 0.25 [38] to 0.5 s [39]. Second most mentioned system, Microsoft Kinect does not allow measurement of APA at all since it is not fitted with any form of pressure measuring equipment. As has been shown in [40] advanced external equipment in form of Zebris FDMT platform was coupled with VR equipment and therefore allowed commencing measurements of better quality, and obtaining data otherwise impossible to gather. This further undermines reasonability of creating and using “Advanced” systems. More examples of such customisation have been shown in [41, 42]. It is becoming apparent, that small, easily accessible and sufficient for benefiting the process in context of results or just introducing the entertainment aspect to rehabilitation process devices are on demand and so far are considered a beneficial addition to rehabilitation process. Based on performed review all authors agree, that said devices are

62

G. Gruszka et al.

developing in direction of miniaturisation and “user-friendliness”. Available components for producing such devices are developing as well, granting progressively more functionalities and possibilities in terms of enhancing physical rehabilitation as time passes and new types of VR devices emerge. Performing this reviews led authors to believe that use of VR is the future of physical rehabilitation, if not for improving the results in a direct manner, then for the aspect of entertainment which tackles the severe problem of unattractiveness of rehabilitation process.

5 Summary Question which arises is if use of expensive and not easily accessible to patients and therapists systems is still viable and provides unique data or allows measurements which would be not possible to perform through use of cheaper systems. Introduction of new functionalities that make systems such as Wii Balance Board or Microsoft Kinect stand up to standards set by professional systems is getting relatively easy— VR sets can be coupled with other, advanced measuring systems, therefore allowing performance of complex measurements in environments mimicking scenarios which would be hard or dangerous to reflect in real world, or would not allow performing intended measurements. Difference in rehabilitation quality is also questionable, due to inaccessibility of the “advanced” systems and impossibility of using them by patients at their homes, while not professional systems are easily accessible and cheaper, taking less space and allowing patients to comfortably train in their homes. On the contrary, high quality components which would allow performing measurements on fast-changing phenomena or replication of complex scenarios are only found in the “advanced” systems. This results in increase in price and what follows - lower accessibility. Therefore, they cannot be reliably used as equipment for completing entire physical rehabilitation process, although they remain valuable devices when it comes to performing research on rehabilitation processes and finding new ways of improving it. Acknowledgements This acknowledgment is only in IiBE template.

References 1. Burdea GC (2003) Virtual rehabilitation - benefits and challenges. Methods Inf Med 42(5):519– 523. https://doi.org/10.1055/s-0038-1634378 2. Nichols PJR, Parish JG (1959) Rehabilitation of fractures of the shafts of the tibia and fibula. Rheumatology 5(3):73–87. https://doi.org/10.1093/rheumatology/5.3.73

Trends in Development ...

63

3. Cano Porras D, Siemonsma P, Inzelberg R, Zeilig G, Plotnik M (2018) Advantages of virtual reality in the rehabilitation of balance and gait: systematic review. Neurology 90(22):1017– 1025. https://doi.org/10.1212/WNL.0000000000005603. 4. Cikajlo I, Matjaçi´c Z (2009) Advantages of virtual reality technology in rehablitation of people with neuromuscular disorders. Recent Adv Biomed Eng. https://doi.org/10.5772/7473 5. Keshner EA (2004) Virtual reality and physical rehabilitation: a new toy or a new research and rehabilitation tool? J Neuroeng Rehabil 1:1–2. https://doi.org/10.1186/1743-0003-1-8 6. Polecho´nski J, Nierwi´nska K, Kalita B, Wodarski P (2020) Can physical activity in immersive virtual reality be attractive and have sufficient intensity to meet health recommendations for obese children? A pilot study. Int J Environ Res Public Health 17(21):1–14. https://doi.org/10. 3390/ijerph17218051 7. Feng H et al (2019) Virtual reality rehabilitation versus conventional physical therapy for improving balance and gait in Parkinson’s disease patients: a randomized controlled trial. Med Sci Monit 25:4186–4192. https://doi.org/10.12659/MSM.916455 8. Dascal J et al (2017) Virtual reality and medical inpatients: a systematic review of randomized, controlled trials. Innov. Clin. Neurosci. 14(1–2):14–21 9. Deutsch JE (2011) Using virtual reality to improve walking post-stroke: translation to individuals with diabetes. J Diabetes Sci Technol 5(2):309–314. https://doi.org/10.1177/ 193229681100500216 10. Kim K, Choi B, Lim W (2019) The efficacy of virtual reality assisted versus traditional rehabilitation intervention on individuals with functional ankle instability: a pilot randomized controlled trial. Disabil Rehabil Assist Technol 14(3):276–280. https://doi.org/10.1080/17483107. 2018.1429501 11. Zhao Y et al (2016) Feasibility of external rhythmic cueing with the Google Glass for improving gait in people with Parkinson’s disease. J Neurol 263(6):1156–1165. https://doi.org/10.1007/ s00415-016-8115-2 12. Howard MC (2017) A meta-analysis and systematic literature review of virtual reality rehabilitation programs. Comput Human Behav 70:317–327. https://doi.org/10.1016/j.chb.2017.01. 013 13. Page MJ et al. (2021) The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 372 (2021). https://doi.org/10.1136/bmj.n71 14. Cano Porras D, Sharon H, Inzelberg R, Ziv-Ner Y, Zeilig G, Plotnik M (2019) Advanced virtual reality-based rehabilitation of balance and gait in clinical practice. Ther Adv Chronic Dis 10:1–16. https://doi.org/10.1177/2040622319868379 15. Ye¸silyaprak SS, Yildirim MS, ¸ Tomruk M, Ertekin Ö, Algun ZC (2016) Comparison of the effects of virtual reality-based balance exercises and conventional exercises on balance and fall risk in older adults living in nursing homes in Turkey. Physiother Theory Pract 32(3):191–201. https://doi.org/10.3109/09593985.2015.1138009 16. Prasertsakul T, Kaimuk P, Chinjenpradit W, Limroongreungrat W, Charoensuk W (2018) The effect of virtual reality-based balance training on motor learning and postural control in healthy adults: a randomized preliminary study. Biomed Eng Online 17(1):1–17. https://doi.org/10. 1186/s12938-018-0550-0 17. Kim EK, Kang JH, Park JS, Jung BH (2012) Clinical feasibility of interactive commercial Nintendo gaming for chronic stroke rehabilitation. J Phys Ther Sci 24(9):901–903. https://doi. org/10.1589/jpts.24.901 18. Cho KH, Lee KJ, Song CH (2012) Virtual-reality balance training with a video-game system improves dynamic balance in chronic stroke patients. Tohoku J Exp Med 228(1):69–74. https:// doi.org/10.1620/tjem.228.69 19. Phu S, Vogrin S, Al Saedi A, Duque G (2019) Balance training using virtual reality improves balance and physical performance in older adults at high risk of falls. Clin Interv Aging 14:1567– 1577 (2019). https://doi.org/10.2147/CIA.S220890. 20. Bateni H (2012) Changes in balance in older adults based on use of physical therapy vs the Wii Fit gaming system: a preliminary study. Physiother (United Kingdom) 98(3):211–216. https:// doi.org/10.1016/j.physio.2011.02.004

64

G. Gruszka et al.

21. Bieryla KA, Dold NM (2013) Feasibility of Wii Fit training to improve clinical measures of balance in older adults. Clin Interv Aging 8:775–781. https://doi.org/10.2147/CIA.S46164 22. Cho GH, Hwangbo G, Shin HS (2014) The effects of virtual reality-based balance training on balance of the elderly. J Phys Ther Sci 26(4):615–617. https://doi.org/10.1589/jpts.26.615 23. Franco JR, Jacobs K, Inzerillo C, Kluzik J (2012) The effect of the Nintendo Wii Fit and exercise in improving balance and quality of life in community dwelling elders. Technol Heal Care 20(2):95–115. https://doi.org/10.3233/THC-2011-0661 24. Jorgensen MG, Laessoe U, Hendriksen C, Nielsen OBF, Aagaard P (2013) Efficacy of Nintendo Wii training on mechanical leg muscle function and postural balance in community-dwelling older adults: a randomized controlled trial. J Gerontol Ser A Biol Sci Med Sci 68(7):845–852. https://doi.org/10.1093/gerona/gls222 25. Yang WC, Wang HK, Wu RM, Lo CS, Lin KH (2016) Home-based virtual reality balance training and conventional balance training in Parkinson’s disease: a randomized controlled trial. J Formos Med Assoc 115(9):734–743. https://doi.org/10.1016/j.jfma.2015.07.012 26. Morone G, et al (2014) The efficacy of balance training with video game-based therapy in subacute stroke patients: a randomized controlled trial. Biomed Res Int 2014. https://doi.org/ 10.1155/2014/580861 27. Cho KH, Lee WH (2014) Effect of treadmill training based real-world video recording on balance and gait in chronic stroke patients: a randomized controlled trial. Gait Posture 39(1):523– 528. https://doi.org/10.1016/j.gaitpost.2013.09.003 28. Cho KH, Lee WH (2013) Virtual walking training program using a real-world video recording for patients with chronic stroke: a pilot study. Am J Phys Med Rehabil 92(5):371–384. https:// doi.org/10.1097/PHM.0b013e31828cd5d3 29. Yang S, Hwang WH, Tsai YC, Liu FK, Hsieh LF, Chern JS (2011) Improving balance skills in patients who had stroke through virtual reality treadmill training. Am J Phys Med Rehabil 90(12):969–978. https://doi.org/10.1097/PHM.0b013e3182389fae 30. Lloréns R, Gil-Gómez JA, Alcañiz M, Colomer C, Noé E (2015) Improvement in balance using a virtual reality-based stepping exercise: a randomized controlled trial involving individuals with chronic stroke. Clin Rehabil 29(3):261–268. https://doi.org/10.1177/0269215514543333 31. Rajaratnam BS et al (2013) Does the inclusion of virtual reality games within conventional rehabilitation enhance balance retraining after a recent episode of stroke? Rehabil Res Pract 2013:1–6. https://doi.org/10.1155/2013/649561 32. Lee HC, Huang CL, Ho SH, Sung WH (2017) The effect of a virtual reality game intervention on balance for patients with stroke: a randomized controlled trial. Games Health J 6(5):303–311. https://doi.org/10.1089/g4h.2016.0109 33. Tsang WWN, Fu ASN (2016) Virtual reality exercise to improve balance control in older adults at risk of falling. Hong Kong Med J 22(1):S19–S22 = Xianggang yi xue za zhi 34. Lloréns R, Noé E, Colomer C, Alcañiz M (2015) Effectiveness, usability, and cost-benefit of a virtual reality-based telerehabilitation program for balance recovery after stroke: a randomized controlled trial. Arch Phys Med Rehabil 96(3):418-425.e2. https://doi.org/10.1016/j. apmr.2014.10.019 35. Lee MM, Lee KJ, Song CH (2018) Game-based virtual reality canoe paddling training to improve postural balance and upper extremity function: a preliminary randomized controlled study of 30 patients with subacute stroke. Med Sci Monit 24:2590–2598. https://doi.org/10. 12659/MSM.906451 36. Chang WD, Chang WY, Lee CL, Feng CY (2013) Validity and reliability of Wii Fit balance board for the assessment of balance of healthy young adults and the elderly. J Phys Ther Sci 25(10):1251–1253. https://doi.org/10.1589/jpts.25.1251 37. Polecho´nski J, Nawrocka A, Wodarski P, Tomik R (2019) Applicability of smartphone for dynamic postural stability evaluation. Biomed Res Int. https://doi.org/10.1155/2019/9753898 38. Gazit E et al (2020) What happens before the first step? A new approach to quantifying gait initiation using a wearable sensor. Gait Posture 76(12):128–135. https://doi.org/10.1016/j. gaitpost.2019.11.007

Trends in Development ...

65

39. Delafontaine A, Vialleron T, Hussein T, Yiou E, Honeine JL, Colnaghi S (2019) Anticipatory postural adjustments during gait initiation in stroke patients. Front Neurol 10. https://doi.org/ 10.3389/fneur.2019.00352 40. Wodarski P, Jurkoj´c J, Chmura M, Bieniek A, Guzik-Kopyto A, Michnik R (2021) Analysis of the ability to maintain the balance of veterans of stabilization missions. In: Innovations in Biomedical Engineering, pp 197–207 41. Wodarski P, Jurkoj´c J, Gzik M (2020) Wavelet decomposition in analysis of impact of virtual reality head mounted display systems on postural stability. Sensors (Switzerland) 20(24):1–12. https://doi.org/10.3390/s20247138 42. Wodarski P, et al. (2020) Assessment of gait stability and preferred walking speed in virtual reality. Acta Bioeng Biomech 22(1). https://doi.org/10.37190/abb-01490-2019-03

Biomechanical Assessment of Selected Motion and Cognitive Exercises in the “Neuroforma” Rehabilitation System Agata Guzik-Kopyto, Katarzyna Nowakowska-Lipiec, Piotr Szaflik, Oliwia Nowicka, and Robert Michnik

Abstract The aim of the study was the biomechanical assessment of selected motor and cognitive exercises carried out with the use of the Neuroforma rehabilitation system. The selected parameters of the kinematics of the upper limbs’ movements during exercises with the Neuroforma system were compared with the kinematics of the movement of selected activities used in daily living. The reference group consisted of 17 adults without any musculoskeletal disorders. For the reference group, studies of the kinematics of the movement of the upper limbs were carried out using the Xsens system. Tasks performed during these activities included actions such as drinking from a cup or lifting an object on a platform. The study of the kinematics of movement of selected cognitive-motor exercises (“Paths”, “Track”) with the Neuroforma system was carried out for one healthy adult. It has been shown that the maximum value of angles and ranges of motion in the shoulder and elbow joints during exercise in the Neuroforma system is higher than in the analyzed everyday activities. Moreover, the values of the analyzed kinematic parameters do not increase linearly with the increase made in the levels of the analyzed exercises. The type and level of exercise in the system should be selected by the doctor/physiotherapist individually on the basis of the patient’s mobility capabilities, based on medical knowledge and a biomechanical assessment of movements made for specific exercises and their levels. Keywords Neuroforma · Cognitive–motor exercises · Motion kinematics · Biomechanics · Virtual reality · Rehabilitation · Upper limb · Activities of daily living

A. Guzik-Kopyto · K. Nowakowska-Lipiec (B) · R. Michnik Department of Biomechatronics, Faculty of Biomedical Engineering, Silesian University of Technology, Roosevelta 40, 41-800 Zabrze, Poland e-mail: [email protected] URL: http://www.ib.polsl.pl P. Szaflik · O. Nowicka Students Scientific “Biokreatywni”, Faculty of Biomedical Engineering, Silesian University of Technology, Roosevelta 40, 41-800 Zabrze, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_7

67

68

A. Guzik-Kopyto et al.

1 Introduction The basic factor which determines a persons’s quality of life is the ability to independently and undisturbedly perform everyday activities. The appearance of motor– organ dysfunction often leads to involuntary exclusion from society, thus lowering an individual’s well-being and independence. The continuous and dynamic development of modern technologies undoubtedly improves a patient’s the quality of life. Continuous development can be also observed in the area of rehabilitation. It is especially appreciated by patients who require intensive rehabilitation in order to recover faster after illnesses and injuries. Virtual reality techniques are increasingly being used in the rehabilitation of patients [4, 6, 7, 13, 16, 18]. Thanks to this, rehabilitation is a more pleasant process, consisting of more than only monotonous exercises. It also stimulates areas of the brain that are not activated during traditional forms of rehabilitation. Virtual reality (VR) is a tool which can be used in rehabilitation to make process more interesting. It is a three-dimensional image that more or less accurately imitates the real world. The image is created using information technology and can show elements of both the real and fictional worlds. This modern technology includes not only visualization, but also sounds, user interactions, smells, and even touch and taste [4]. It can be observed that the topic of using VR in rehabilitation is being discussed more and more frequently, [4] e.g. with the use of the Kinect motion sensor [1, 5, 14], VR goggles [18], and VR caves [3, 7]. Thanks to the use of virtual reality technology that allows the user to interact with the console through the use of gestures of the body’s limbs, it is possible to rehabilitate a wide range of movement dysfunctions, injuries, and disorders. During training with the use of virtual reality, the user can be scanned into the 3D space, thanks to which it is possible to assess the patient’s condition and ability to perform exercises [1, 14]. Virtual rehabilitation is increasingly used among patients. Research indicates that it may be helpful in patients with balance disorders [4], patients with limb paresis [13], or with the elderly as a means of supporting the process of rehabilitation [6]. One of the types of rehabilitation in virtual reality with the use of biofeedback is the Neuroforma system [19]. It is intended for independent movement and cognitive exercises. It is used by patients with neurological diseases and injuries. The Neuroforma system has access to 20 exercise modules. They simultaneously engage motor and cognitive functions, thereby improving the precision of movements, eye-hand coordination, joint mobility, muscle strength, and endurance, as well as the processes of perception, decision-making, attention, and memory, among others [19]. Rehabilitation exercises in the world of virtual reality should be designed on the basis of medical knowledge and knowledge of biomechanics of the musculoskeletal system, e.g. knowledge of the course of angles in the joints over time and knowledge of the ranges of motion for a given exercise/daily activity, both for healthy people and people with motor deficits. The creators may base their research on the research of the kinematics of motion. Many publications present research into the kinematics

Biomechanical Assessment of Selected Motion and Cognitive Exercises ...

69

of everyday activities, such as e.g. drinking from a cup [2, 8, 11], washing the face [2], eating with a spoon, or brushing hair [2, 12]. The aim of the study was the biomechanical assessment of selected motor and cognitive exercises as performed with the use of Neuroforma rehabilitation equipment. It was decided to compare the selected parameters of the kinematics of the upper limb’s movement during exercises with the Neuroforma system with the kinematics of the movement of selected activities of daily living.

2 Materials and Methods The reference group consisted of 17 healthy people: men (age: 22.76 ± 2.39 years, body height: 1.80 ± 0.07 m, body weight: 74.53 ± 8.9 kg, BMI: 23.03 ± 2.44). The dominant limb in all persons was the right limb. The kinematics of the movement of the upper limb during the performance of two everyday activities (drinking from a cup and lifting an object) was investigated. The measurement conditions were the same for the whole group, and each activity was performed by the examined person in three consecutive trials [8]. The MVN Biomech system by Xsens with built-in IMU sensors was used to conduct the research. Additionally, studies of the kinematics of the movement of the upper limbs were carried out during exercises on the Neuroforma rehabilitation equipment Meden-Inmed for one person without disorders within the musculoskeletal system (male, age: 21 years, body height: 1.78 m, body weight: 67 kg, dominant upper limb: left) [19]. Motion kinematics studies were also carried out using the Xsens system. During the examination, the person was positioned in front of a screen where he saw his real, mirror image, around which virtual objects appeared (Fig. 1a, b). The task of the examined person was to direct his reflection in such a way as to grab and move the appearing object along the outline of a given shape. Research with the use of Neuroforma was carried out on various levels of difficulty (level 4, 11, 18, and 25 out of 29 possible) by performing exercises from various categories. Three trials were performed for each level. After each test, the examined person had time to rest. Two exercises were analyzed: • exercise 1 (“Paths”)—consisted in moving a series of balls along the displayed lines. As the difficulty level increased, the balls were moved asymmetrically and at different speeds (Fig. 1a), • exercise 2 (“Track”)—consisted in grasping car with one hand and moving the object along the displayed track. As the difficulty level increased, the occurrence of changes in the driving speed increased and the track became more and more complicated. The person exercising during this task used only one upper limb—the research was conducted for the dominant limb (Fig. 2b). The analysis covered the maximum angular values and ranges of movements: abduction-adduction, internal-external rotation, flexion-extension in the shoulder joint, and flexion-extension in the elbow joint. All movements were recorded during

70

A. Guzik-Kopyto et al.

Fig. 1 The person examined during performing: a exercise 1 “Paths”, b exercise 2 “Track”

exercises with the Neuroforma system. It was decided to examine how the abovementioned parameters changed in the analysed exercised successive levels of difficulty. The obtained results of the above-mentioned parameters for exercises with the Neuroforma system were compared with the values obtained during everyday activities.

3 Results The obtained results, i.e. the maximum angular values and ranges of movements: abduction-adduction, internal-external rotation, flexion-extension in the shoulder joint, and flexion-extension in the elbow joint, for exercises with the Neuroforma system and everyday activities, are presented in Figs. 2 and 3. For the exercises with Neuroforma, the graphs show the mean values with the standard deviation from 3 trials for each difficulty level and the reference group—the mean value with the deviation obtained for the entire group.

Biomechanical Assessment of Selected Motion and Cognitive Exercises ...

71

Fig. 2 The obtained results: abduction of the shoulder joint of a the right upper limb, b the left upper limb; rotation of the shoulder joint of c the right upper limb, d the left upper limb

72

A. Guzik-Kopyto et al.

Fig. 3 The obtained results: flexion of the shoulder joint of a the right upper limb, b the left upper limb; flexion of the elbow joint of c the right upper limb, d the left upper limb

4 Discussion Virtual rehabilitation refers to the use of virtual reality elements in the rehabilitation process. It has been proven that this method rehabilitation leads to the improvement of postural stability in people with labyrinth dysfunction, improves balance, supports the control of body posture, reduces the risk of falling [4], improves the mobility of the paresis of the upper limb in patients after an ischemic stroke [1, 13–15], leads to an increase in muscle strength in the rehabilitated limbs [6, 13], helps in the recovery of locomotor functions in patients after a stroke, improves gait function [4, 9], and supports the process of rehabilitation of patients over 65 years of age [6]. The aim of this study was the biomechanical evaluation of selected motorcognitive exercises as conducted with the use of the rehabilitation system Neuroforma. It was decided to compare selected parameters of the kinematics of the move-

Biomechanical Assessment of Selected Motion and Cognitive Exercises ...

73

ment of the upper limb during exercises performed using the Neuroforma system with the kinematics of the movement of selected activities of daily living. By analyzing the results of kinematics for selected activities of daily living, i.e. drinking from a cup or lifting, it was found that the analyzed ranges of motion are similar to the values obtained by the authors of other works [2, 8, 11, 12]. While conducting a comparative analysis of the results, it was noted that when performing exercises with the Neuroforma system, the maximum values of angles and ranges of motion in the shoulder and elbow joints were higher than for the analyzed activities of daily living (Figs. 2 and 3). For exercises in the Neuroforma system, the ranges of movements of the dominant limb for subsequent difficulty levels 4/11/18/25 were, respectively: • for adduction-abduction in the shoulder joint for exercise 1: 46.36◦ ± 1.79◦ /56.75◦ ± 4.90◦ /50.77◦ ± 0.28◦ /61.43◦ ± 2.85◦ , for exercise 2: 54.55◦ ± 5.2◦ /54.70◦ ± 4.28◦ /53.62◦ ± 2.18◦ /61.21◦ ± 1.13◦ ; • for rotation in the shoulder for exercise 1: 69.7◦ ± 6.53◦ /93.36◦ ± 5.58◦ /93.68◦ ± 8.4◦ /104.33◦ ± 9.09◦ , for exercise 2: 71.31◦ ± 7.11◦ /92.69◦ ± 3.08◦ /89.36◦ ± 0.99◦ /101.55◦ ± 4.02◦ ; • for flexion in the shoulder joint for exercise 1: 85.74◦ ± 6.79◦ /98.55◦ ± 3.81◦ / 107.07◦ ± 4.72◦ /104.66◦ ± 8.17◦ , for exercise 2: 85.23◦ ± 1.9◦ /93.69◦ ± 3.52◦ / 95.49◦ ± 3.39◦ /96.62◦ ± 1.31◦ ; • for bending in the elbow joint for exercise 1: 92.76◦ ± 9.06◦ /87.3◦ ± 3.67◦ / 87.43◦ ± 11.14◦ /100.8◦ ± 6.05◦ , for exercise 2: 112.64◦ ± 11.64◦ /83.04◦ ± 6.16◦ /88.57◦ ± 16.9◦ /104.12◦ ± 12.59◦ ; For the abduction and rotation movement in the shoulder joint, the maximum values of angles in the joints measured in exercise 1 “Paths” were almost twice as high as for activities of daily living. For exercise 1 in the Neuroforma system, higher values of the maximum flexion angles in the shoulder joint were also noted (85.19◦ ± 5.53◦ /101.42◦ ± 4.88◦ /103.99◦ ± 3.27◦ /108.22◦ ± 8.86◦ ) with regard to selected activities of everyday life, while for the exercise “Track” these values are similar (they assume values of approx 89.73◦ ± 2.53◦ /100.4◦ ± 4.07◦ /99.95◦ ± 1.21◦ /103.75◦ ± 2.77◦ ). The maximum values of the flexion angle in the elbow joint are higher for the analyzed activities of daily living. During exercises of the dominant limb in the Neuroforma system they are: 96.44◦ ± 6.98◦ /94.31◦ ± 4.46◦ /93.09◦ ± 9.2◦ /105.63◦ ± 2.19◦ for the subsequent analyzed levels of exercise 1 and 108.16◦ ± 15.14◦ /82.4◦ ± 3.82◦ /89.44◦ ± 9.66◦ /105.21◦ ± 12.38◦ for the subsequent analyzed levels of exercise 2. When using the Neuroforma system for rehabilitation purposes, the degree of motor deficit should be taken into account, as people with more severe disabilities may not be able to perform the exercise even at a low level of difficulty. Moreover, it was noted that the maximum values of angles and the ranges of angles in the joints did not increase linearly with the growing levels. The higher levels of exercise 1 “Paths” makes the balls move asymmetrically at different speeds, and in exercise 2, the path on which the object should be led assumes a more complicated shape. However,

74

A. Guzik-Kopyto et al.

for both exercises, the range of motion and the maximum angle in the joint do not change (i.e. do not intensify) as the difficulty level increases. This is quite important information for a physiotherapist who conducts an exercise with Neuroforma system and who must select the appropriate level of exercise and type of exercise for a specific patient with paresis of the upper limbs. However, the Neuroforma system can be used in the rehabilitation of cognitive processes and patients with impaired sensor and motor integration. The exercises are based on visual perception, visual-auditory coordination, and the synchronization of movements. Most of the studies presented in the literature discuss the subject of patients’ efficiency after rehabilitation, in which VR techniques are used [1, 4, 6, 9, 13, 14]. Few papers discuss the analysis of the kinematics of the movement of the upper limbs during exercises in virtual rehabilitation [10, 17]. There are still no scientific reports on the individualization of rehabilitation exercises in VR based on the patient’s mobility (biomechanics of movement). Before the rehabilitation system is approved for general use, it should undergo a series of biomechanical tests in terms of its applicability to patients with specific motor injuries/deficits. The programs used during the rehabilitation process must be able to individually select rehabilitation exercises for specific cases and ensure the safety of patients during use. Limitations and Directions of Further Research The authors are aware of the presented research’s limitations. Only one person was subjected to the study of exercises in the Neuroforma system. In further stages of the work, the group of participants should be expanded. Two selected cognitivemotor exercises were analyzed. The research can also be extended to other types of exercises. The direction of further analyzes may also include the assessment of postural stability during exercise, as well as the assessment of well-being and the physiological parameters of using the Neuroforma system for a longer period of time.

5 Conclusions Based on the analysis of the obtained results, the following conclusions were formulated: • the maximum values of angles and ranges of motion in the shoulder and elbow joints during exercise with the Neuroforma system are higher than found with the analyzed activities of daily living, i.e. drinking from a cup and lifting an object on a platform, • the maximum values of angles and ranges of angles in the shoulder and elbow joints do not increase linearly with the increase in the levels of the analyzed exercises of the Neuroforma system, • the type and level of exercise must be selected by the doctor/physiotherapist individually, according to the mobility of the patient with paresis of the upper limbs. The doctor/physiotherapist should relay on medical knowledge and a biomechanical assessment of movements for specific exercises and their levels.

Biomechanical Assessment of Selected Motion and Cognitive Exercises ...

75

References 1. Adomaviˇcien˙e A, Daunoraviˇcien˙e K, Kubilius R, Varžaityt˙e L, Raistenskis J (2019) Influence of new technologies on post-stroke rehabilitation: a comparison of Armeo spring to the kinect system. Medicina (Kaunas) 55(4):98 2. Aizawa J, Masuda T, Koyama T, Nakamaru K, Isozaki K, Okawa A, Morita S (2010) Threedimensional motion of the upper extremity joints during various activities of daily living. J Biomech 43:2915–2922 3. Borrego A, Latorre J, Llorens R, Alcañiz M, Noé E (2016) Feasibility of a walking virtual reality system for rehabilitation: objective and subjective parameters. J Neuroeng Rehabil 13(1):68 4. Cano Porras D, Siemonsma P, Inzelberg R, Zeilig G, Plotnik M (2018) Advantages of virtual reality in the rehabilitation of balance and gait: systematic review. Neurology 90(22):1017– 1025 5. Chanpimol S, Seamon B, Hernandez H, Harris-Love M, Blackman MR (2017) Using Xbox kinect motion capture technology to improve clinical rehabilitation outcomes for balance and cardiovascular health in an individual with chronic TBI. Arch physiother 7:6 6. de Amorim J, Leite RC, Brizola R, Yonamine CY (2018) Virtual reality therapy for rehabilitation of balance in the elderly: a systematic review and META-analysis. Adv Rheumatol (London, England) 58(1):18 7. Gieser S, Becker E, Makedon F (2013) Using CAVE in physical rehabilitation exercises for rheumatoid arthritis. In: ACM international conference proceeding series. https://doi.org/10. 1145/2504335.2504367 8. Guzik-Kopyto A, Michnik R, Wodarski P, Chuchnowska I (2016) Determination of loads in the joints of the upper limb during activities of daily living. In: Pie˛tka E, Badura P, Kawa J, Wieclawek W (eds) Information technologies in medicine 5th international conference ITIB ´ ˛ski, Poland, 20–22 June 2016. Advances in Intelligent Systems and Com2016, Kamie´n Sla puting, vol 472. Springer, Cham, pp 99–108 9. Harkema SJ, Schmidt-Read M, Lorenz DJ, Edgerton VR, Behrman AL (2012) Balance and ambulation improvements in individuals with chronic incomplete spinal cord injury using locomotor training-based rehabilitation. Arch Phys Med Rehabil 93(9):1508–1517 10. Hussain N, Murphy MA, Sunnerhagen KS (2018) Upper limb kinematics in stroke and healthy controls using target-to-target task in virtual reality. Front Neurol 9:300 11. Kyung K, Won-Kyung S, Jeongsu L, Hwi-Young L, Dae-Sung P, Byung-Woo K, Jongbae K (2014) Kinematic analysis of upper extremity movement during drinking in hemiplegic subjects. Clin Biomech 29:248–256 12. Magermans DJ, Chadwick EKJ, Veeger HEJ, van der Helm FCT (2005) Requirements for upper extremity motions during activities of daily living. Clin Biomech 20:591–599 13. Mekbib DB, Han J, Zhang L, Fang S, Jiang H, Zhu J, Roe AW, Xu D (2020) Virtual reality therapy for upper limb rehabilitation in patients with stroke: a meta-analysis of randomized clinical trials. Brain Inj 34(4):456–465 14. Park DS, Lee DG, Lee K, Lee G (2017) Effects of virtual reality training using Xbox kinect on motor function in stroke survivors: a preliminary study. J Stroke Cerebrovasc Diseases Off J Natl Stroke Assoc 26(10):2313–2319 15. Stryla W, Banas A (2016) The use of virtual reality technologies during physiotherapy of the paretic upper limb in patients after ischemic stroke. J Neurol Neurosci 6:3 16. Sveistrup H (2004) Motor rehabilitation using virtual reality. J Neuroeng Rehabil 1(1):10 17. Trincado-Alonso F, Dimbwadyo-Terrer I, de los Reyes-Guzmán A, López-Monteagudo P, Bernal-Sahún A, Gil-Agudo Á (2014) Kinematic metrics based on the virtual reality system Toyra as an assessment of the upper limb rehabilitation in people with spinal cord injury. BioMed Res Int 904985 (2014) 18. Wodarski P, Jurkoj´c J, Polecho´nski J, Bieniek A, Chrzan M, Michnik R, Gzik M (2020) Assessment of gait stability and preferred walking speed in virtual reality. Acta Bioeng Biomech 22(1):127–134 19. https://www.neuroforma.pl/

Thermovision-Based Human Body Temperature Measurement Supported by Vision System Damian Krawczyk, Paulina So´sniak, Weronika Czech, Michał Swierzy, Kajetan Łado´s, Łukasz Seweryn, Michał Zwardon, ´ Sławomir Suchon, ´ Wojciech Wolanski, ´ Rafał Setlak, and Ziemowit Ostrowski

Abstract The article presents the concept of using thermal image processing to measure temperature, but with the support of a classic vision system and the digital image processing and recognition. As part of the research, the hybrid thermovision-vision system was built, the purpose of which was to search for characteristic measurement areas on the human face that are reliable for the measurement of body temperature. The research focused on measurements in the corners of the eyes. The selection of the measurement area was based on the analysis and recognition of visual images, while the temperature was determined on the basis of the analysis of the infrared image of the studied area. Research was carried out on a small research group, the results were compared with those obtained with the use of non-contact medical thermometers. The obtained results, after taking into account the conditions in which the experiments were carried out, can be regarded as satisfactory and confirming the validity of the adopted concept of a hybrid temperature measurement system. Keywords Thermovision · Body temperature · Vision systems · Covid-19

1 Introduction Fast, reliable and minimally invasive measurement of human body temperature has gained importance in the recent period associated with the spread of the Covid-19 coronavirus pandemic. Temperature measurement allows for quick identification of people potentially infected with this virus by detecting elevated body temperature. The technology that makes it possible to perform this type of measurements is thermography. D. Krawczyk · M. Swierzy · K. Łado´s · Ł. Seweryn · M. Zwardo´n · R. Setlak · Z. Ostrowski Department of Mechatronics, Faculty of Electrical Engineering, Silesian University of Technology, Gliwice, Poland e-mail: [email protected] URL: http://mechatronika.polsl.pl/index.php P. So´sniak · W. Czech · S. Sucho´n (B) · W. Wola´nski Department of Biomechatronics, Faculty of Biomedical Engineering, Silesian University of Technology, Zabrze, Poland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_8

77

78

D. Krawczyk et al.

The use of thermovision in measuring the temperature of the human body is a fairly widely described [4, 8, 12]. In [4] the authors use the data obtained from thermography at the airport to assess passenger health questionnaires. Mostly, these methods require the recognition of the face, which is the area of measurement, at some stage. For face recognition, the authors use a thermal image [8] or an image from a coupled visible light or IR camera [12]. Considerations on integrating temperature measurement with a building automation system can be found e.g. in works [5]. A fairly extensive exploration of the topic of using a thermal imaging camera for temperature measurement in farm animals is presented in [9, 15]. The aim of the work described in the article is to develop an automated system for measuring the temperature of the human body. The system will be based on a thermal imaging camera and images provided by a visible light camera. The system is to be integrated with BMS (Building Management System) or constitute its module. The expected result of presented algorithm will be to increase the effectiveness of protection against COVID-19. The architecture of the developed system will enable its integration in rooms equipped with building automation systems without significant interference. With the above in mind, the system can be quickly deployed in buildings, limiting the spread of biological hazards.

2 Methodology 2.1 Selecting the Measurement Area The normal, internal, human body temperature is stated as 36.4–37 ◦ C [7]. Internal measurement is most precise for determination of body temperature, however from practical point of view it is cumbersome. For this reason more common method is non-invasive measurement even taking into account disadvantages. The non-invasive method of temperature measurement is associated with less accuracy measurement related to the influence of external conditions. Temperature difference internal, and the skin temperature for a healthy person may be up to 17 ◦ C [2]. Also the ambient temperature is one of factors determining measurement accuracy. Depending on the ambient temperature and the part of the body where the measurement is made, fluctuations of up to 14.8 can be expected [1]. One way of non-invasive measurement is thermovision. The area selected for temperature measurement using thermovision must be as accessible as possible to the camera and the lowest possible susceptibility to differences caused by external factors. The forehead (easily accessible) can be a reasonable compromise due to the ease of detection by the facial recognition algorithm, however, it is exposed to various factors that may lower or increase the temperature (wind, sun, sweat, etc.) and is discouraged [3]. Forehead as region of interest was used in [8, 12]. The best point with the given characteristics is the lacrimal muscle (Latin. lacrimal caruncle) Fig. 1. This area is located inside the face, minimizing the impact of wind and outside

Thermovision ...

79

Fig. 1 Anatomy of human eye, lacrimal caruncle visible in left eye corner [17]

temperature, while its strong blood supply allows for an accurate assessment of internal temperature [6]. On the other hand, it is a small point which is definitely a challenge. Lacrimal muscle is also used for measuring the body temperature of the horses [9]. The temperature of the lacrimal muscle for a healthy person is 34.8 ◦ C (32.2– 36.0 ◦ C) [11]. Deviation beyond this value is associated with fever or cold the body.

2.2 Test Stand The Flir A325 thermal imaging camera and the classic Samsung SNB-6004P visible light camera were selected for the purpose of the measurements. Testing stand including both cameras should allow to measure temperature of selected area of lacrimal muscle. Using vision systems with Samsung camera area of interest was selected, then measurement of temperature were taken using Flir camera. Due to the need to overlay the field of view of both cameras, a test stand was designed to enable proper positioning of both cameras. For this purpose, a photographic tripod was used with elements produced by 3D printing Fig. 2. Fusion of vision and thermal imaging in medicine has found application for example in diagnostic of skin [10]. Room temperature during test was 20 ◦ C and humidity was 60%. Distance between face and cameras was 2 m.

80

D. Krawczyk et al.

Fig. 2 Set-up of testing stand with cameras

2.3 Face Recognition and Temperature Measurement Algorithm The general idea of algorithm can be summarised in three steps: 1. 2. 3. 4.

obtaining visual and thermal image, recognition of the eye corner using visual image, fusion of visual image and thermal image, reading temperature of the eye corner using thermal image.

The Matlab environment was used to connect to the vision camera using Image Acquisition Toolbox [16]. It allows you to connect to any IP camera, with Real Time Streaming Protocol (RSTP) support, providing image encoded using the H.264 standard. The connection is established after creation the ipcam object. Additionally, it is possible to provide appropriate data to authenticate the connection with the camera and the time after which the connection attempt is ended. Another tool that speaks in favor of using this toolbox is the function of capturing images using the appropriate command sent by Matlab to the camera. It allows for quick and efficient preparation of data for processing. In the research plan, Matlab Image processing Toolbox was also to be used to obtain a thermal image, however due to unsolved

Thermovision ...

81

problems with compatibility, ThermaCAM Researcher Pro 2.9 was used to obtain the images from Flir camera. The most important part of a developing the algorithm is recognition of the face and its characteristic elements including the corner of the eye, selected as measurement point. The Matlab Computer Vision Toolbox were used. Implementation used in the project is based on the Viola-Jones’s algorithm [14]. The algorithm consists of three main steps: – selection of Haar’s features – creating an integral image – creating and training the Haar’s classifier [13]. The detection of the characteristic elements of the face is determine by the likelihood of them occurrence by calculating the difference in pixel intensity. An integral image can be defined as a matrix containing the sum of the pixels intensities directly to the left and directly above the pixel with coordinates (x, y). The integral image (AI ) is created from original image (A) according to the formula (Eq. 1): AI [x, y] =



A(x  , y  )

(1)

x  ≤x,y  ≤y

A tool for affine matrix transformation was used to merge the images, thus allowing for linear image mapping. User on starting the algorithm selects points in both images that show the same element such as a marker on a wall or other characteristic points. Using this input the fusion of images is made. After this operations, temperature value of eye corner is extracted from thermal image.

2.4 Implementation to Nazca BMS Developed algorithm was introduce to Nazca Building Management System— (BMS). Nazca is a universal automation platform developed by APA Group, allowing the management of intelligent building control systems. Nazca is used by Mechatronics and Biomechatronics departments in students laboratories which allows practical application of developed solution. First step of implementation of face recognition and temperature measurement algorithm to BMS was establish communication between Nazca and Matlab. For this purpose the customizable node in Nazca was prepared. The customizable node allows on any configuration using the code written in C# language. The measurement data is exchanged between Nazca and Matlab by saving and reading data from a csv file.

82

D. Krawczyk et al.

Fig. 3 Nazca GUI presenting graph of the measured temperature with archival measurements and the current temperature reading with communicate. (top-left corner title is “chart”, top-right corner title is “temperature”, bottom-right communicate on green field indicates “Temperature is low”) (Color figure online)

The next step is visualization of the measurement results by the operator station (graphic user interface of Nazca) Fig. 3. OP presents a graph of the measured temperature with archival measurements and the current temperature reading with one of the corresponding communicate: – “Temperature is low” (for readings below 36,6 ◦ C) – “Temperature is elevated” (for readings 36.6–37 ◦ C) – “Temperature is high” (for readings above 37 ◦ C)

3 Results Algorithm was tested using two reference devices: Microlife NC 150 and Medivon Timi Verona handheld non-contact thermometers. Measurements were made on the group of students participating in the project and project supervisors (7 person total). After acclimatization in room, each tested person underwent a triple temperature measurement with both non-contact thermometers, and the obtained results were averaged. Then at each the person was measured using a thermal imaging camera mounted on the designed test stand. For best accuracy face measuring was made on low-reflecting background, with minimal air flow and away from potential sources of IR radiation such as room radiators. Collected measurement are presented in Table 1.

Thermovision ...

83

Table 1 Measurements for 7 person using non-contact medical thermometer (average value from 3 measurements) and using Flir A325 camera (without and with ATC correction), last four columns present difference of thermovision system vs medical thermometers [◦ C] Person Microlife NC 150 avg std

Medivon Timi Verona avg std

Flir A325 Flir A325 with Difference ATC corr.

Difference with ATC corr.

1 2 3 4 5 6 7

36.43 36.10 36.53 36.17 37.13 36.63 36.13

35.1 34.9 34.8 35 35.4 35.2 34.7

0.21 0.29 0.79 0.19 0.08 −0.49 −0.51 0.08

36.03 35.90 36.30 35.90 36.20 35.43 34.90

0.31 0.10 0.17 0 0.10 0.42 0.26

0.15 0 0.06 0.06 0.12 0.06 0.06

35.82 35.61 35.51 35.71 36.12 35.92 35.41 Avg difference:

0.93 1.00 1.50 0.90 0.80 0.23 0.20 1.11

1.33 1.20 1.73 1.17 1.73 1.43 1.43 1.43

0.61 0.49 1.02 0.46 1.01 0.71 0.72 0.72

Fig. 4 Average body temperature measured for research group using non-contact thermometers and developed algorithm

4 Discussion Differences between thermovision body measurements and verification medical thermometers are 1.11 and 1.43 ◦ C in average and 1.73 ◦ C max for person 3 and 5. It can be observed that measurements using Flir A325 are lowered. The reason of this trend most probably is disregarding of Atmospheric Transmission Constant (ATC). Atmosphere between measuring device and object partially block radiation emitted by object. To correct this ATC can be calculated and introduced to measuring method. ATC was calculated using Flir software [18] and is 0.98. After taking into account ATC correction for Flir A325 average difference drops to 0.08 and 0.72 and 1.02 ◦ C for person 3. Considering that average error (thermovision

84

D. Krawczyk et al.

vs verification thermometers) is comparative to standard deviation from 3 measurements using medical thermometers, accuracy achieved is satisfactory (Fig. 4). Research group is to small to draw significant conclusion however observed trends are promising. Further research on larger and more varied group are required to formulate binding conclusions. The final conclusion is that thermovision can be used for screening human body temperature measurements as part of smart home systems. Advantages of cooperating of thermal imagining and vision image processing allows to achieve precise location of measurement which is a necessary condition for methodology. Among other advantages possibility of automation and identity recognition can be indicated. Disadvantages are mainly connected to general requirements of non-internal temperature measurement such as: acclimatization, environmental factors and individual conditioning. Acknowledgements Presented materials are results of Individual Study Programs implemented in the form of Project Based Learning—Edition no 5 as part of the project: “Silesian University of Technology as a Center of Modern Education based on research and innovation”, PBL project title: “Thermovision measurement of human body temperature assisted by digital image processing”.

References 1. Beker BM, Cervellera C, De Vito A, Musso CG (2018) Human physiology in extreme heat and cold. Int Arch Clin Physiol 1:001. https://doi.org/10.23937/iacph-2017/1710001 2. Chan LS, Lo JL, Kumana CR, Cheung BM (2013) Utility of infrared thermography for screening febrile subjects. Hong Kong Med J 19(2):109–115. PMID: 23535669 3. Cheung BM, Chan LS, Lauder IJ, Kumana CR (2012) Detection of body temperature with infrared thermography: accuracy in detection of fever. Hong Kong Med J 18(Suppl 3):31–34. PMID: 22865221 4. Cho KS, Yoon J (2014) Fever screening and detection of febrile arrivals at an international airport in Korea: association among self-reported fever, infrared thermal camera scanning, and tympanic temperature. Epidemiol Health 36:e2014004. https://doi.org/10.4178/epih/ e2014004. Accessed 30 May 5. Christopoulos K, Antonopoulos C, Voros N, Orfanoudakis, T (2017) Building automation systems in the world of internet of things. In: Keramidas G, Voros N, Hübner M (eds) Components and services for IoT platforms. Springer, Cham. https://doi.org/10.1007/978-3-31942304-318 6. Huggins J, Rakobowchuk M (2019) Utility of lacrimal caruncle infrared thermography when monitoring alterations in autonomic activity in healthy humans. Eur J Appl Physiol 119(2):531– 538. https://doi.org/10.1007/s00421-018-4041-6. Epub 4 Dec 2018. PMID: 30515591 7. Hutchison JS et al (2008) Hypothermia therapy after traumatic brain injury in children. New Engl J Med. 358 (23):2447–2456. https://doi.org/10.1056/NEJMoa0706930. PMID 18525042. S2CID 46833 8. Jia-Wei L, Ming-Hung L, Yuan-Hsiang L (2019) A thermal camera based continuous body temperature measurement system. In: Proceedings of the IEEE/CVF international conference on computer vision (ICCV) 9. Kim SM, Cho GJ (2021) Validation of eye temperature assessed using infrared thermography as an indicator of welfare in horses. Appl Sci 11:7186. https://doi.org/10.3390/app11167186

Thermovision ...

85

10. Ledwo´n D, Juszczyk J, Pie˛tka E (2019) Infrared and visible image fusion objective evaluation method. In: International conference on information technologies in biomedicine. Springer, Cham 11. Mokrzecka J, Pietruszynska M, Jankowska I, Grabska-Liberek I (2017) Termografia w okulistyce - pomiar temperatury powierzchni oka, Borgis - Post¸epy Nauk Medycznych 3/2017, pp 135–139 12. Sharma A, Yadav AR (2017) Image processing based body temperature estimation using thermal video sequence. In: International conference on computing methodologies and communication (ICCMC), pp 846–852. https://doi.org/10.1109/ICCMC.2017.8282585 13. Wilson PI, Fernandez JD (2006) Facial feature detection using Haar classifiers. J Comput Sci Coll 21(2006):127–133 14. Viola P, Jones M (2001) Rapid object detection using a boosted cascade of simple features. In: Proceedings of the 2001 IEEE computer society conference on computer vision and pattern recognition, CVPR, p I. https://doi.org/10.1109/CVPR.2001.990517 15. Zaiqin Z, Hang Z, Tonghai L (2019) Study on body temperature detection of pig based on infrared technology: a review. In: Artificial intelligence in agriculture, vol 1, pp 14–26. https:// doi.org/10.1016/j.aiia.2019.02.002. ISSN 2589-7217 16. https://www.mathworks.com/help/supportpkg/ipcamera/ug/ipcam.html. Accessed 06 June 2021 17. https://upload.wikimedia.org/wikipedia/commons/a/a9/Gray892.png. Accessed 28 Jan 2021 18. https://flir.custhelp.com/app/answers/detail/a-id/2989/~/how-do-i-calculate-theatmospheric-transmission-constant. Accessed 06 June 2021

Critical Analysis of Recreational Activities as a Method to Reduce Obesity Małgorzata Matyja, Joanna Szołtysek, and Andrzej W. Mitas

Abstract The article examines a controversial problem of criticising physical activity in the aspect of physiological irregularities or incorrect performance of motoric activities. The authors represent a questionable view that a wrong physical activity is actually worse than no activity at all. They formulate the following thesis: walking with poles (Nordic Walking) is bad for obese people and undermines the basic principles of obesity treatment. The article presents biochemical and anthropometric characteristics of obesity. Main problems related to obesity and musculoskeletal system are discussed in the context of spinal joints and lower limbs overload and degeneration (specifically knee joints). The article examines biomechanical aspects of multiple musculoskeletal loads in obese persons, which lead to dynamic stability disorders. Keywords Obesity · Physical activity · Physiotherapy in obesity · Recreation in obesity

1 Introduction Poor level of physical culture, resulting from, inter alia, faulty educational model, in which physical activity is regarded as a school burden, while sick notes become standard solutions applied by ‘precautious’ parents, implicates developments in healthrelated attitudes and behaviours. Physical activity in a broader sense is not yet a

M. Matyja Department of Physiotherapy, Jerzy Kukuczka University of Physical Education in Katowice, Katowice, Poland J. Szołtysek Department of Rehabilitation, School of Health Sciences, Medical University of Silesia, Katowice, Poland A. W. Mitas (B) Faculty of Biomedical Engineering, Silesian University of Technology, Zabrze, Poland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_9

87

88

M. Matyja et al.

fashion trend, as we are still catching up with the earlier stage of civilizational development, enjoying our car rides as close to destination as possible. After we complete our education (earlier or later, as the case may be), we tend to give in to the mentally inspiring professional absorption and overlook progressive obesity, as small changes cumulate and become problematic only after exceeding the tolerance level. The same happens in typical analogue circuits injuries, i.e. parametric injuries that may sometimes promote catastrophic injuries (myocardial infarction or stroke?). Once the tolerance threshold is exceeded (e.g. when one breathes heavily when trying tie up shoes), an internal conviction appears that it is necessary to act. Astoundingly, we expect to lose weight fast, over a few weeks, although we took years to gain it. In the context of omnipresent instantness, we are looking for an antidote to quick fix intrinsic conditions. In the context of ubiquitous immediacy, we are looking for an immediate antidote to spontaneously occurring conditions (excessive weight appears on its own, while weight reduction requires structured approach). Disappointed with inefficacy of internet-sold pills, we discover that physical activity is the best way to reduce body mass. This faulty way of thinking is due to “skipping” physics and chemistry classes at school. Children should learn about energy balance as early as primary school but apparently this is not so, as the myth of miraculous effects of physical activity for weight loss is not dispelled. Just a reminder: a glass of a sweetened carbonated beverage and a chocolate doughnut is ca. 530 kcal. Three or five minutes of pleasure requires at least an hour of high impact activity. As counterbalance we quote from the people who deal with physical activity: ‘. . . Swimming and physical exercise are great contributors to slimming. In water, the body becomes lighter and we can perform movements which are difficult to make in a natural environment. Water produces high resistance, which requires a lot of power to overcome as well as precise movements. That is why you do not develop high speed when swimming, but you work on your muscular power and shape your body in a harmonious way. . .’ Faith works miracles, but you cannot lose weight this way, if just because muscles (provided you develop then) are heavier than fat. One must not overlook the comment that it is easier to perform movements in water - in the light of the principle of maximum entropy it will have just one effect in a statistical majority of cases - the movements performed will be detrimental and aggravating the acquired disfunctions, being simply easier and painless. A genuine common problem is the fact that it is easy to introduce energy in a chemical form to the human body while it is difficult to take it out as work and heat. Low-impact walking with poles seem to be a right way to lose weight and one that is particularly interesting during isolation periods. After a brief training (a couple of hours at most) and disregarding boundary conditions, such as structural limitations or general lack of knowledge about the way human body functions, allured by the apparent ease of Nordic Walking, we embark on this activity. For people in the X range of age, e.g. for the people who come up with numerous excuses (laziness?) not to take up activities that require diligence, training or non-speedy gain of experience, walking becomes one of a few accessible types of physical activity. Walking popularity will grow, while obesity is already a wide-spread social phenomenon.

Critical Analysis of Recreational Activities as a Method to Reduce Obesity

89

Before starting this recreational activity, it is worthwhile to consider its biomechatronic aspect. Biomedical engineers should make sure to inform the publicity about its easily predictable ‘side effects’, which can be analysed with the use or applicable models even before they appear and cause damage that is difficult to repair. A simple example: it is hard to imagine one may walk comfortably with a bag of cement suspended to one’s waist at the front, and this is what weight distribution may look like in obese walkers.

2 Motivation and Description of the Problem Correct body structure is the initial condition for each physical activity, including Nordic Walking, if the goal function of the activity is recreation. Each physical activity has its own specific rules, and failure to meet them leads to negative consequences. A different type of activity is rehabilitation, whereby the goal function is the correct performance of a task that activates selected body parts. Poor physical condition (e.g. due to obesity) accompanied by incorrect physical activities are the worst option possible. Unfortunately, it is a very frequent one. There are many people who walk with hip flexion and distinctive upper body part support on a vertical pole positioned away from the body. This is just one of the ways to inflict self-harm distributed over time. In the context of the main limiting factor for a full-swing workout, i.e. obesity, the authors discussed briefly its treatment options and reviewed recommended dietary regimes, focusing mainly on a limited catalogue of allowed physical activities. Special attention should be paid to a brief critical review of bibliography related to recreational activities recommended to the obese (recreation cannot play the role of therapeutic measures) - sports, jogging, gym. In literature, the authors usually focus on the effects on behavioural and biochemical parameters while overlooking the activity’s consequences for the human locomotor system. Nordic Walking is often perceived as an instant remedy for mobility atrophy. Review of literature on positive effects of NW (body weight loss, improved blood biochemical parameters) with disregard to critical approach to a specific status of mass overloaded musculoskeletal system of the obese patients is usually the domain of NW enthusiasts, who claim that a few house of training is enough to reap its benefits.

3 Basic Medical and Biomechanical Problems of Obesity Literature Review A contemporary person with a sedentary lifestyle is burning ca. 300 kcal on physical activity per day while consuming ca. 2100 kcal daily in meals, a proportion of 1:7.

90

M. Matyja et al.

According to Saris, in order to ‘revert’ to the energy balance of our predecessors, a contemporary man should increase their daily energy expenditure by ca. 400 kcal, an equivalent of 45–60 min of aerobic exercise per day [13]. Obese people are faster to develop diabetes and atherosclerosis and have higher prevalence of hypertension, ischemic heart disease, cardiovascular failure, stroke, hyperlipidaemia as well as some musculoskeletal overloads and tumours. The higher the body mass, the bigger the pressure on joints. Obese people struggle with lower extremities joints overload, particularly in the knee joint, which performs rotation, flexion and extension. A joint with significant overload is slowly losing its functions. That is why many obese people are not able to fully extend or flex their knees. Obesity also causes hip and feet joint overload. With higher load on joints caused by excess weight, the risk of injuries goes up by almost 80%. Joint disfunctions and degeneration are related to excess body weight but also to specific effects of adipose tissue [9]. Excess fat tissue plays an important role in the pathogenesis of osteoarthritis. Obesity is related to modified hormonal and growth factor production, which can accelerate chondroblast division [7, 10]. Also, obesity is a chronic inflammatory process, in which excess proinflammatory agents are produced by the white adipose tissue, such as the Interleukin-1, Interleukin-6, TNF-α (tumour necrosis factor α), as well as adipokines: leptin, adiponectin, resistin and visfatin [7]. Adipokines, i.e. fat tissue hormones take part in many metabolic processes related to glucose or lipid metabolism and they modulate immunological and inflammatory response [3, 6, 7, 17]. Adipokines damage joint tissues, i.e. cartilage, synovial membrane and bones. In the light of the above, it should be emphasized that excess weight and obesity promote degenerative disease not only within the joints with excess weight overload, such as knee, joint and ankle but also in upper extremities joints [6, 14]. In biomechanics, it is assumed that body weight and height and the area of support determine stability of the standing posture. Hence a hypothesis that the higher the body mass, the more stable the standing position. Dynamic stability represents a different problem. An obese person needs to engage much more muscular power to regain stability. Obesity-associated changes decrease the muscular system to body mass ratio and thus it is more difficult for an obese person to restore body balance. Increased body weight disables activation of postural equilibrium reactions and impairs dynamic corrections that restore body balance. Literature shows that stability workout may only be started in overweight patients after their body mass reduction [2].

4 Body Mass Reduction Methods In a majority of cases, obesity treatment should be based on diet modification and increased physical activity, with development of new behaviour patterns accompanied by psychological and pharmacological support, and in case of morbid obesity (BMI > 35 kg/m2 , and particularly > 40 kg/m2 ) the treatment potentially could

Critical Analysis of Recreational Activities as a Method to Reduce Obesity

91

include surgery [11]. A weight losing diet should be composed to include possibly all groups of food (cereals, dairy, meat and fish and their products, fats, fruit and vegetables) to allow provision of necessary nutrients. The diet should be well balanced to maintain coherent meal structure, frequency and regularity (5 times per day). Therefore, it is important not to skip morning meals, not to snack in between meals, not to overeat at dinner, which should be consumed ca. 3–4 h before bedtime. Meals should be eaten slowly as the feeling of satiety only occurs after some time from starting the meal. The basic feature of dietary regimes is long-term increase of fat consumption accompanied by higher intake of vegetables, fruits, cereal and, usually, regular physical activity [16]. Physical activity is a very important issue. The choice of physical activity types should be individualized and preceded by the assessment of the condition of the large joints, obesity degree and the patient’s history in physical exercise. Given the load on lower extremities and spine in overweight patients (BMI > 35), the first recommended activity is water exercise and cycling. These are the safest forms of physical workout for overweight people as they cause no overload on their lower limbs’ joint surfaces. It should be emphasized that it is important to individualize also these types of physical exercise, which are dedicated to obese patients, e.g. cycling. In this type of exercise, the patient’s body weight is supported by a bicycle, thus protecting ankle, knee and hip joints from harmful load. Cycling requires a correct (straight) body posture, which can be achieved by regulating the handlebar and the bicycle seat. Seat height positioning also impacts the degree of knee flexion. To ensure comfortable workout, the knee flexion should be ca. 5 with the patient’s foot at its lowest position. Higher knee flexion produces excessive knee compression when the patient’s foot is at its top point during pedalling. With frequent and long-term repetition, it can contribute to exacerbation or generate knee pain [12].

4.1 Critical Introspection of the Effects of Excessive Physical Activity in the Obese Patients In the light of the above-presented principles for treating obese patients, it is worthwhile to have a critical look into such application of Nordic Walking. This type of recreation cannot be treated as therapy. In literature, body weight reduction associated with NW in obese patients is described as follows: ‘. . . four-week Nordic Walking training programme brought about increased exercise tolerance, body weight reduction and lowering of cardiovascular metabolic risk factors. . . ’ or ‘. . . after a 10-week training, the NW group recorded average reduction in body mass (6.4%), BMI (6.4%), blood glucose (3.8%), total cholesterol (10.4%), LDL cholesterol (12.8%), triglycerides (10.6%) and increase in HDL cholesterol (9.6%)....” [5]. Certainly, when an obese person starts to engage in more physical activity, they can lose some extra kilos. However, these types of publications that focus on biochemical and anthropometric factors fail to account for the potential negative effects of NW on the musculoskeletal system.

92

M. Matyja et al.

In this context it is worthwhile to mention publications that discuss the negative effects of NW, in which the authors focused on NW impact on the human musculoskeletal system. The studies conducted by German orthopaedists demonstrate that NW produces a higher load on the knee joint immediately after the initial contact of the heel with the ground than the natural gait. This is due to longer strides and bigger foot position angle. It was also concluded that although NW group recorded smaller torques in sagittal and frontal planes in the knee joints as compared to the runners, but in the transverse plane the torques were higher than in spontaneous gait or run. On the basis of these findings, the authors suggest that NW should not be recommended as physical exercise for the people who wish to reduce biomechanical load on their lower extremities [15]. Although other studies recorded lower mechanical load of the foot during forward push as compared to natural gait (lower peak vertical force), they also demonstrated higher horizontal and vertical forces at foot landing, which indicates higher mechanical load on the foot. Therefore, it is doubtful if NW should continue to be recommended to overweight patients and patients with lower limb musculoskeletal problems [8]. It was also concluded that NW training produces increased ground reaction forces on the heel during swing phase, particularly during brisk NW. Given these findings, the authors revisit their opinion about smaller load on individual foot parts during this type of physical activity. Such loads may in fact increase and result in feet deformations. The researchers also believe that distorted propulsion triggers spine joints overload [1]. Taking into account the arguments presented above, one may conclude that in case of obesity, the adverse effects of NW are multiplied as compared to its detrimental impact for people with correct body weight, and therefore NW activity should not be recommended to obese patients. It seems that many publications present an oversimplified view on NW in the obese patients, focusing on potential weight loss and improvement in blood biochemical parameters that accompany obesity, while disregarding NW negative consequences for the musculoskeletal system.

4.2 Physical Activity as Motivating and Rehabilitating Factor Public opinion regards physical activity as a vital health factor. But health improvement is not the only motivation for physical activity. There are a few types of motivation: health and hygienic type - in which physical activity is undertaken to keep fit and maintain perfect silhouette; an activity-hedonistic type, whereby people simply feel the need to exercise; a relax-catharsis type related to the need to evacuate stress and negative emotions; explorer type and ambitious type characterize people who seek adrenaline-rush activities that provide specific type of risk and possibility to test one’s limits; cognitive and educational type characterizes people who drive towards self-improvement and wish to develop their passions; a social type, whereby physical activity is a perfect opportunity for meeting other people [4].

Critical Analysis of Recreational Activities as a Method to Reduce Obesity

93

It should be added that the type of physical activity recommended for the people with diseases should be aligned to their disfunctions, regardless of their motivation type.

4.3 Conclusions and Findings The article discusses recreational character of NW in the context of its recommendation for the obese patients by physiotherapy professionals and in the aspect of the view that: 1. recreation cannot be considered as therapy; 2. patients should first lose weight and then engage in rehabilitation workouts, e.g. for body stability, stronger muscles and efficiency. The articles comments on the oversimplified approach to the application of NW in obesity but such critical review should also be conducted for other diseases and disfunctions. Reference should be made to the musculoskeletal system parameters and not just the parameters characteristic for individual diseases. Correct walking with poles requires a long, ca. 8-week training in order to master the right walking style. This requirement is not emphasized in literature. In the context of the general lack of time, it is wrong to promote the popular NW as the ideal type of physical activity, which is uncritically applied in both healthy and sick patents and leads to complications in disfunctions.

References 1. Cie´sla W, Gieremek K, Drabik J, Górny M (2015) Analiza pedobarograficzna obci¸az˙ enia stopy w czasie chodu z u˙zyciem kijów nordic walking. Medycyna Sportowa 31(3):129–135 ´ 2. Cie´sli´nska-Swider J (2004) Stabilno´sc´ posturalna otyłych kobiet przed i po kuracji odchudzaja˛cej [PhD dissertation]. AWF Katowice 3. Conde J, Scotece M, Gomez R, Lopez V, Gomez-Reino JJ, Gualillo O (2011) Adipokines and osteoarthritis: novel molecules involved in the pathogenesis and progression of disease. Arthritis 2011 4. Góreczna A, Garczy´nski W (2017) Motivations for physical activity-literature review. J Educ Health Sport 7(7):322–337 5. Hagner-Derengowska M et al (2015) Effects of Nordic walking and Pilates exercise programs on blood glucose and lipid profile in overweight and obese postmenopausal women in an experimental, nonrandomized, open-label, prospective controlled trial. Menopause 22(11):1215– 1223 6. Issa R, Griffin T (2012) Pathobiology of obesity and osteoarthritis: integrating biomechanics and inflammation. Pathobiol. Aging Age-related Dis. 2(1):17470 7. King LK, March L, Anandacoomarasamy A (2013) Obesity & osteoarthritis. Indian J Med Res 138(2):185

94

M. Matyja et al.

8. Kleindienst FI, Michel KJ, Schwarz J, Krabbe B (2006) Comparison of kinematic and kinetic parameters between the locomotion patterns in nordic walking, walking and running. Sportverletzung Sportschaden: Organ der Gesellschaft fur Orthopadisch-Traumatologische Sportmedizin 20(1), 25 9. Koszowska A, Nowak J, Hawranek R (2015) Choroba zwyrodnieniowa stawów w kontek´scie nadwagi i otyło´sci. In: Forum Zaburze´n Metabolicznych, vol 6, no 2, pp 56–63 10. Lementowski PW, Zelicof SB (2008) Obesity and osteoarthritis. Am. J. Orthopedics-Belle Mead- 37(3):148 11. Ostrowska L (2010) Leczenie dietetyczne otyło´sci-wskazówki dla lekarzy praktyków. In: Forum Zaburze´n Metabolicznych, vol 1, no 1, pp 22–30 12. Plewa M, Markiewicz A (2006) Aktywno´sc´ fizyczna w profilaktyce i leczeniu otyło´sci. Endokrynologia, Otyło´sc´ i Zaburzenia Przemiany Materii 2(1):30–37 13. Saris WHM et al (2003) How much physical activity is enough to prevent unhealthy weight gain? Outcome of the IASO 1st Stock Conference and consensus statement. Obes Rev 4(2):101– 114 14. Stanisławska-Biernat E (2008) Społeczne i ekonomiczne aspekty choroby zwyrodnieniowej stawów. Pol Arch Med Wewn 118:50–53 15. Stief F, Kleindienst FI, Wiemeyer J, Wedel F, Campe S, Krabbe B (2008) Inverse dynamic analysis of the lower extremities during nordic walking, walking, and running. J Appl Biomech 24(4):351–359 16. Tsigos C et al (2008) Management of obesity in adults: European clinical practice guidelines. Obes Facts 1(2):106–116 17. Vincent HK, Heywood K, Connelly J, Hurley RW (2012) Obesity and weight loss in the treatment and prevention of osteoarthritis. PM&R 4(5):S59–S67

Assessment of the Time of Electromechanical Muscle Response to a Given Rhythmic Sound Stimulus Robert Michnik, Aneta Danecka, Anna Manka, ´ and Andrzej W. Mitas

Abstract The aim of the study was to assess the electromechanical response times of muscles and to assess the effect of repeatability of the performed task on the indicated times: TRT (total reaction time), PMT (premotor time), and EMD (electromechanical delay). The test consisted in squeezing the dynamometer with maximum force in the shortest possible time after hearing an audible signal that repeated 15 times at 5-second intervals. The analysis clearly shows that there is an influence of the repeatability of the performed task on the designated reaction times after the occurrence of the sound stimulus. It was noticed that the applied rhythmic sound stimulations result in a shortening of the premotor reaction time. Keywords Electromechanical delay · Premotor time · Response times · Rhythmic auditory stimulation

1 Introduction How quickly a human muscle is able to respond to a given stimulus? The so-called electromechanical response times are quite accurate in answering these questions. The term electromechanical response times consists of premotor time (PMT) and electromechanical delay time (EMD), which are collectively referred to in the literature as total reaction time (TRT). Electromechanical response times can be confessed by using EMG [12, 15, 17]. TRT is the time measured from the occurrence of a given stimulus to the beginning of the moment of force generation by muscles. PMT is defined as the time from the stimulus to the beginning of bioelectrical activity of the muscle, whereas EMD is the time from the beginning of the electrical activity of the muscle to the beginning of its strength generation [12, 15]. The premotor response time measures the central nervous system latency and includes the time that takes into account the detection of the stimulus, its transmission and integration of the stimulus in the central nervous system and the processing of relevant movement commands. R. Michnik (B) · A. Danecka · A. Ma´nka · A. W. Mitas Faculty of Biomedical Engineering, Silesian University of Technology, Zabrze, Poland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_10

95

96

R. Michnik et al.

EMD measures peripheral nervous system latency and is considered a valuable tool for assessing neuromuscular coordination [6, 17].

1.1 Total Response Time - TRT The length of the total response time to a stimulus depends on a number of factors that additionally influence its duration. Scientific reports indicate that this time depends on: the type of stimulus, intensity of the stimulus, muscle stimulation, concentration state of the subject, age of the patient, and gender [1, 3, 4]. The research has shown that men in practically every age group have a shorter response time to any stimulus than women [1, 3]. In addition, it has been noted that for both women and men, the increase in response time with age is similar [3]. The analysis of the effect of age on response time showed a downward trend in the age of 0–20 and an upward trend after 50 years of age in both genders [3, 4].

1.2 Electromechanical Delay Time - EMD Scientists state that the EMD value is mainly influenced by the susceptibility of the muscle tendon, which results in the explanation of the electromechanical delay of the muscle. The value of EMD in individual studies, however, varies - for human skeletal muscles this value ranges from 30 to even several hundred milliseconds depending on the source. Analyses of EMD time confirm that it depends on the type and structure of the skeletal muscles studied, their fatigue or the type of work undertaken [2, 5, 14, 16].

1.3 Rhythmic Auditory Stimulation Electromechanical response times can be examined using visual or auditory stimuli, which are intended to indicate the moment when muscles need to be stimulated. It is stated that on a sound stimulus the human response is faster than on a visual stimulus - the difference is due to the difference in time needed for the central nervous system to reach the appropriate stimuli [7]. In rehabilitation, the use of sound stimuli, which are aimed at working out correct patterns of rhythmic movement is referred to as Rhythmic Auditory Stimulation (RAS). RAS provides some kind of guidance for human body movement, and the effectiveness of this method has been confirmed by recent studies, where the results provide clear evidence of the beneficial effect of RAS on the temporal-spatial parameters of gait in dysfunctional and healthy individuals [8, 10, 11].

Assessment of the Time of Electromechanical Muscle ...

97

1.4 Aim of the Study The aim of the study was to assess the electro-mechanical response times of the muscles responsible for palm squeeze (long palm muscle and finger flexor) and to assess the effect of the task’s repeatability on the determined TRT, PMT, and EMD times. On the basis of the literature, a hypothesis was formulated that the electromechanical responses of the examined muscles with the next test will be shortened and the electromechanical delay time will take a constant value. Rhythmic sound stimuli were used in the study to develop the technique of individualisation of sound rhythm adjustment.

2 Methodology The research group consisted of 30 people, with an average age of 23 2, where 77% were women. 93% of the respondents were right-handed. To record the data, a measurement system for EMG tests—Noraxon TeleMyo2400 G2 was used, along with a pressure dynamometer and computer software MyoResearch XP Master Edition 1.07.01. In order to perform the measurement, each examined person on the dominant forearm in the area of the muscles responsible for the hand squeeze (musculus palmaris longus, musculus flexor digitorum profundus) was glued with surface EMG electrodes, which are to determine the moment of activation of the upper limb muscle group. In addition, the measuring station was equipped with document-type headphones. The test consisted of squeezing the dynamometer with maximum force in the shortest possible time after hearing an acoustic signal which was repeated 15 times at 5-second intervals. Additionally, 2 s after the first signal, a command appeared which determined the moment when the person performing the exercise could release the hug of the measuring device. The EMG signals were additionally subjected to standard processing, where the ratification and signal smoothing operations were performed. The treatment of the raw EMG signal is aimed at increasing the accuracy and reliability of the obtained information. In the first stage, a signal cleaning operation was performed, where all negative amplitudes are reflected by the base line in the positive direction. In the second stage a digital signal smoothing algorithm was used - moving average algorithm, which emphasizes the main direction of the EMG signal [9]. The digital signal processing procedures used may have little impact on the time measurements applied. In order to avoid such effects, only operations that do not significantly change the average shape of the curves have been used. Additionally, no cut-off filters were used (e.g. cutting out 50 60 Hz noise) because too much power of EMG signal is lost [9]. The obtained records were further analysed in the Matlab software, where an algorithm was created, which allowed for automatic determination of the beginnings of recording activity for individual muscle groups and for the dynamometer. Automatic detection of the beginning of muscle activity

98

R. Michnik et al.

Fig. 1 Test runs with automatic detection (red markers - sound initiation, red round marker - start of muscle activation/on dynamometer)

was implemented using the method of calculating the range of standard deviations for the baseline EMG, each time before the beginning of a given muscle activity for 1 s. For the conducted tests a multiplier was established, that in case of 5% exceeding the SD range, the moment of muscle group activation is determined. In order to eliminate the treatment of single discharges as the moment of activation, a time of minimum time of the subperiod (MTS) was determined, during which the EMG signal should constantly remain above the threshold value. In the cases analysed, the MTS time has been set at 500 ms and the threshold value at 25% of the standard deviation range exceeded. Verification of the reliability of the established threshold values for the conducted tests was performed using graphical methods - Fig. 1 [9]. In an analogous way, threshold conditions were established for the value of the hand tightening force. On the basis of the knowledge of the moment when the sound signal appears (every 5 s), individual times of electromechanical muscle response to a given stimulus were determined: premotor response time - the time between the sound signal and the activity of muscle groups, and electromechanical delay - the time between the muscular activity and the beginning of the recording of the value of hand tightness. The total reaction time of a human to a sound stimulus was determined as the sum of both these times. Graphic interpretation of individual times is presented in Fig. 2.

Assessment of the Time of Electromechanical Muscle ...

99

Fig. 2 Graphical representation of electromechanical response times

Table 1 Average values of electromechanical response times of muscles responsible for squeezing

Mean ± SD [s] (95% CI) CV [%]

TRT

PMT

EMD

0,34 ± 0,16 (0,26 – 0,42) 46,10

0,18 ± 0,15 (0,10 – 0,26) 85,86

0,16 ± 0,05 (0,14 – 0,19) 29,44

3 Results The analysis of the obtained electromechanical response times determined the averages together with standard deviations for the relevant components for the whole test group - Table 1. Due to the expected high standard deviations obtained for the calculated averages, it was decided to divide the study into 3 parts and compare the parameters for the start of the study (sample 1–5), the middle of the study (6–10 samples) and the end of the study (11–15 samples). The obtained averages with deviations are shown in Fig. 3. For the total response time and for the premotor response of the muscle with subsequent parts of the test, we notice a decrease in the duration of these parameters. This decrease for TRT is estimated at 19 between the beginning and the centre of the test and 4% between the centre and the end of the test. For PMT the percentage decrease is 33% and 11%. In the case of EMD, it can be seen that this parameter oscillates at a constant value and the differences between the individual samples do not exceed 2%. To examine the effect of the repeatability of the task on the determined TRT, PMT and EMD times, the mean values obtained by the test group separately for each sample with standard deviations are shown in the Table 2 and in the bar graphs Fig. 4. Analysing the obtained distributions for individual samples, it can be seen that the total muscle response time to the stimulus and the premotor response time record a decrease in values with the next sample. For TRT, there is a 3% decrease, while for PMT, there is a 4% decrease. No correlation is recorded for electromechanical delay. In addition, a statistical analysis was performed to determine the statistical significance of the results obtained. In the first stage, the Shapiro-Wilk normality test was performed. If the test was statistically significant (p < 0.05), it indicated a distribution distant from the Gaussian curve. For the obtained results, normal distribution

100

R. Michnik et al.

Table 2 Average values of electromechanical response times of muscles responsible for squeezing together with standard deviation and variability within the group, relative to the sample number TRT

PMT

EMD

Mean ± SD CV [%] (95% CI) [s]

Mean ± SD CV [%] (95% CI) [s]

Mean ± SD CV [%] (95% CI) [s]

1

0,48 ± 0,19 39,39 (0,41 – 0,55)

0,32 ± 0,18 56,93 (0,25 – 0,57)

0,16 ± 0,08 47,85 (0,14 – 0,30)

2

0,44 ± 0,18 40,10 (0,38 – 0,51)

0,29 ± 0,17 59,31 (0,23 – 0,52)

0,15 ± 0,05 32,33 (0,14 – 0,29)

3

0,37 ± 0,19 (0,31–0,44)

51,11

0,21 ± 0,18 82,62 (0,15 – 0,36)

0,16 ± 0,04 27,27 (0,15 – 0,31)

4

0,36 ± 0,12 34,37 (0,31 – 0,40)

0,20 ± 0,11 54,51 (0,16 – 0,37)

0,15 ± 0,05 34,87 (0,13 – 0,29)

5

0,33 ± 0,14 42,49 (0,28 – 0,38)

0,17 ± 0,14 84,93 (0,12 – 0,28)

0,16 ± 0,05 33,13 (0,14 – 0,31)

6

0,32 ± 0,17 52,21 (0,26 – 0,38)

0,14 ± 0,15 109,09 (0,09 – 0,23)

0,18 ± 0,06 31,20 (0,16 – 0,34)

7

0,35 ± 0,11 32,02 (0,31 – 0,39)

0,20 ± 0,11 52,76 (0,17 – 0,37)

0,15 ± 0,05 33,04 (0,13 – 0,28)

8

0,30 ± 0,12 39,61 (0,26 – 0,34)

0,14 ± 0,13 90,73 (0,09 – 0,23)

0,16 ± 0,03 17,98 (0,15 – 0,31)

9

0,28 ± 0,08 28,64 (0,26 – 0,31)

0,12 ± 0,06 50,81 (0,10 – 0,22)

0,16 ± 0,04 26,65 (0,15 – 0,31)

10

0,35 ± 0,22 63,22 (0,27 – 0,43)

0,19 ± 0,23 120,23 (0,11 – 0,30)

0,16 ± 0,04 24,83 (0,15 – 0,31)

11

0,32 ± 0,15 46,97 (0,27 – 0,38)

0,15 ± 0,25 99,44 (0,10 – 0,25)

0,17 ± 0,04 23,03 (0,15 – 0,32)

12

0,32 ± 0,16 50,68 (0,26 – 0,38)

0,16 ± 0,16 101,11 (0,10 – 0,26)

0,16 ± 0,04 24,75 (0,15 – 0,31)

13

0,32 ± 0,11 34,39 (0,28 – 0,36)

0,16 ± 0,11 69,48 (0,12 – 0,28)

0,17 ± 0,03 19,63 (0,15 – 0,32)

14

0,29 ± 0,09 29,89 (0,26 – 0,32)

0,12 ± 0,09 74,35 (0,09 – 0,21)

0,17 ± 0,04 25,90 (0,15 – 0,32)

15

0,28 ± 0,14 50,10 (0,23 – 0,33)

0,12 ± 0,13 108,19 (0,07 – 0,19)

0,16 ± 0,05 30,45 (0,14 – 0,30)

was recorded for 47% of sample sets, therefore, for further assessment of statistical significance, a non-parametric statistical test - the Wilcoxon test - was used. As part of this study, the assessment of statistical significance of statistical differences was carried out for each subsequent sample by assessing it against the first. For the TRT and PMT times from sample 3 to the last, statistical differences were found to be statistically significant. No statistically significant differences were observed for EMD electromechanical delay.

Assessment of the Time of Electromechanical Muscle ...

101

Fig. 3 Average electromechanical muscle response times for the different parts of the test

Fig. 4 Average a total response time, b premotor response time, c electromechanical delay time, relative to the test number

4 Conclusion The analysis of the results shows that there is an impact of the repeatability of the task on the response times after the sound stimulus. Analysing the PMT, which takes into account many processes related to the detection, transmission of stimuli in the central nervous system, it can be concluded that the cyclic rhythm of activation of these processes may shorten the TRT to the stimulus. In most cases, the research group examined made an attempt to predict when the next stimulus will occur, which means that the human brain is able to learn to react to the given stimulus more quickly. The lack of significantly statistical differences in electromechanical delay times obtained is a satisfactory result. According to the literature review, an extension of the EMD time was expected due to muscle strain fatigue caused by cyclic training during the study [13]. As a result of the study, it was noted that the applied RAS (every 5 s) results in a reduction of the PMT. In the future, it is recommended to conduct similar studies, where the applied sounds will be given at shorter intervals - e.g., every 2 s.

102

R. Michnik et al.

References 1. Brown AA, Derkyi-Kwarteng L, Ackom CK, Addae E, Amoah FND, Blemano DN (2017) Simple reaction time: how it relates to body mass index (BMI), gender and handedness in Ghanaian students. J Med Sci 49(1) 2. Cavanagh PR, Komi PV (1979) Electromechanical delay in human skeletal muscle under concentric and eccentric contractions. Eur J Appl Physiol Occup Physiol 42(3):159–163 3. Der G, Deary IJ (2006) Age and sex differences in reaction time in adulthood: results from the United Kingdom Health and Lifestyle Survey. Psychol Aging 21(1):62 4. Geis CE (2010) Factors affecting reaction time. Sci J Rev 1969:1–23 5. Hamill J, Knutzen KM (2006) Biomechanical basis of Human Movement. Lippincott Williams & Wilkins, Philadelphia 6. Kaneko F, Onari K, Kawaguchi K, Tsukisaka K, Roy SH (2002) Electromechanical delay after ACL reconstruction: an innovative method for investigating central and peripheral contributions. J Orthop Sports Phys Therapy 32(4):158–165 7. Kemp BJ (1973) Reaction time of young and elderly subjects in relation to perceptual deprivation and signal-on versus signal-off conditions. Dev Psychol 8(2):268 8. Konieczna-Nowak L (ed.) (2016) Music Therapy and Emotional Expression: A Kaleidoscope of Perspectives. Karol Szymanowski Academy of Music, Katowice 9. Konrad P (2007) ABC EMG: praktyczne wprowadzenie do elektromiografii kinezjologicznej. Technomex Spółka z o.o 10. Lee S, Lee K, Song C (2018) Gait training with bilateral rhythmic auditory stimulation in stroke patients: a randomized controlled trial. Brain Sci 8(9):164 11. Murgia M et al (2018) The use of footstep sounds as rhythmic auditory stimulation for gait rehabilitation in Parkinson’s disease: a randomized controlled trial. Front Neurol 9, 348 12. Shultz SJ, Perrin DH (1999) Using surface electromyography to assess sex differences in neuromuscular response characteristics. J Athl Train 34(2):165 13. Szpala A, Rutkowska-Kucharska A, Drapala J (2014) Electromechanical delay of abdominal muscles is modified by low back pain prevention exercise. Acta Bioeng Biomech 16(3) 14. Szpala A, Rutkowska-Kucharska A (2017) Electromechanical response times in the knee muscles in young and old women. Muscle Nerve 56(6):E147–E153 15. Winter EM, Brookes FBC (1991) Electromechanical response times and muscle elasticity in men and women. Eur J Appl Physiol Occup Physiol 63(2):124–128 16. Yavuz SU, ¸ Sendemir-Ürkmez ¸ A, Türker KS (2010) Effect of gender, age, fatigue and contraction level on electromechanical delay. Clin Neurophysiolo 121(10):1700–1706 17. Yeung SS, Au AL, Chow CC (1999) Effects of fatigue on the temporal neuromuscular control of vastus medialis muscle in humans. Eur J Appl Physiol Occup Physiol 80(4):379–385

Is the Coronavirus Pandemic Going to ‘Kill’ the Physical Activity of Young People? Robert Michnik, Katarzyna Nowakowska-Lipiec, Katarzyna Jochymczyk-Wo´zniak, Aneta Danecka, Karolina Mika, and Hanna Zadon´

Abstract This work aimed to determine the impact of the restrictions imposed during the SARS-CoV-2 coronavirus pandemic on the physical activity of young people who were staying in the territory of Poland. The investigations included 11 subjects at the age of 20–25. The measurements consisted in the monitoring of a daily number of steps using a mobile device with the installed application Samsung Health or Health and notes about additional physical activities. The data had already been being collected for 7 weeks before the pandemic was announced in Poland and then during the pandemic for a period of 10 weeks. The analysis encompassed a weekly: number of steps, duration of physical activities, number of burnt kcal and number of METs. The COVID-19 epidemic resulted in a ‘dangerous’ drop in physical activity. In the first weeks of the lockdown, an average daily number of steps equalled approximately 1500. Keywords COVID-19 · Lockdown · MET · SARS-CoV-2 · Number of steps

1 Introduction First cases of atypical pneumonia were observed in November 2019, in the city of Wuhan, China. That led to the identification of a new virus from the family of coronaviruses - SARS-CoV-2 - responsible for causing the disease called COVID-19. As a result of an increasing number of cases of this disease outside China - on 11 March 2020, the World Health Organization (WHO) decided to declare the outbreak of the world pandemic [17]. The coronavirus pandemic is a serious health crisis affecting dwellers of all continents, except Antarctica. So far, i.e. till the end of December R. Michnik · K. Nowakowska-Lipiec (B) · K. Jochymczyk-Wo´zniak · A. Danecka · H. Zado´n Department of Biomechatronics, Faculty of Biomedical Engineering, Silesian University of Technology, Roosevelta 40, 41-800 Zabrze, Poland e-mail: [email protected] K. Mika Students’ Scientific Circle “Biokreatywni”, Faculty of Biomedical Engineering, Silesian University of Technology, Roosevelta 40, 41-800 Zabrze, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_11

103

104

R. Michnik et al.

2020, there were over 81 million cases and over 1.81 million of confirmed deaths [18]. Spreading all over the world, the coronavirus pandemic forced the authorities of different states to undertake actions to restrict the range of the COVD-19 outbreak. As a result of the world health crisis, most governments declared the state of emergency and introduced orders and recommendations of social distancing and isolation. The coronavirus pandemic and related lockdown restrictions completely stopped or greatly reduced our natural physical activity resulting from a normal lifestyle. The lockdown entailed the limitation of movement, closure of parks, forests and gyms. That became a problem for people who actively did sports and also people whose physical activity resulted from the movement in connection with work, school or other duties. Initial data confirm that there occurred a dramatic decline in physical activity from the beginning of the year to 22 March 2020. In Europe, there was a reduction - from 7 to 38%—of a number of steps in comparison with the data from previous year [19]. Immobility caused by lockdown and social distancing, lack of physical activity and a sedentary lifestyle may negatively influence both physical and mental condition. A prolonged period of the lack of physical activity in society may lead to: arterial hypertension [15], osteoporosis, malignant neoplasm of breast [13] and prostate, heart and vascular diseases [16], depression [16], backbone pain and muscular pain [6], obesity, diabetes as well as weakening of intellectual functions [16]. On the other hand, psychological stress connected with isolation leads to endocrinological changes, which in turn have impact on immunity and contribute to occurrence of anxiety, symptoms of post-traumatic stress disorder or anger [3]. The key to counteract the outcome of the sedentary lifestyle is to keep up physical activity, its most frequent form being walking [5]. The World Health Organization— WHO points out that during one week adults should do at least 150 min of moderate physical activity or 75 min of intense physical activity; or the sum of moderate and intense physical activities should equal at least 600 MET-min [20]. This work aimed to determine the impact of imposed lockdown restrictions during the SARS-CoV-2 coronavirus pandemic on physical activity of people aged 20–25. These investigations refer to persons staying in the territory of Poland.

2 Materials and Methods The investigations encompassed 11 subjects (students), 8 women and 3 men. The average age of the subjects was 22 ± 1.5 years old, the average body mass equalled 64 ± 14.2 kg and the average height was 168.2 ± 8.2 cm. The tests consisted in the monitoring of an everyday number of steps by means of a mobile device including the application of Samsung Health, or in the case of the IOS system - the Health application. The test participants declared that they were going to use the abovementioned applications all the time during their movement in and outside home. Moreover, any declared additional physical activity done by each person was written down throughout the day, including the time necessary for its execution. The analyses took into consideration only the activities which lasted longer than 10 min. The data

Is the Coronavirus Pandemic Going to ‘Kill’ the Physical Activity of Young People?

105

were gathered from 27 January to 25 May 2020 and divided into two periods. The first one encompassed 7 full week cycles before the announcement of the COVID19 epidemic in Poland, whereas the second one included 10 week cycles during the lockdown. Each weekly number of steps and declared additional physical activity were calculated into the metabolic equivalent of task (MET) in accordance with the Compendium of Physical Activities [21]. The following items were subjected to analysis: weekly number of steps [steps], weekly duration of physical activities (duration of gait and other declared activities) [min], weekly number of burnt kcal due to executed steps and additional physical activities [kcal],weekly number of METs [MET-min/week]. The obtained results were subjected to the statistical analysis. Variable quantities of the analyzed parameters were described by means of a mean value, standard deviation as well as a minimum and maximum value. Normality of the distribution of analyzed variables was checked using the Shapiro-Wilk test. The differences between values of analyzed parameters in subsequent weeks were checked by ANOVA tests for dependent variables or ANOVA Friedman. When significant differences were discovered, the post-hoc tests Tukey HSD or Conover-Inman were carried out. The performed statistical analyses adopted a level of significance p = 0.05. Computations involved software programme PQStat 1.8 manufactured by the PQStat Software company.

3 Results Table 1 presents the obtained mean values as well as minimum and maximum weekly number of steps and the duration of physical activities recorded before the pandemic (Weeks 1–7) and during COVID-19 (Weeks 8–17) in Poland. Table 2 shows the comparison of the obtained results of a weekly number of kcal and weekly METs. The most often recorded additional physical activities included: exercises at home, workout in the gym, yoga and cycling. Out of 11 participants, 4 persons regularly did some additional physical activities, whereas 7 persons sporadically did some or none. Figure 1 presents a weekly number of steps, whereas Fig. 2 - a weekly number of METs before and during the epidemic of COVID-19 for each tested subject (point diagrams). A linear diagram of the grey colour joins weekly mean values of the number of steps/METs for the whole group before the epidemic, while the orange diagram shows the values after the epidemic. In the above-mentioned diagrams, letters A, B, C, D, E, F denote the time of the introduction of subsequent restrictions in Poland or the moment of lifting the imposed restrictions; the details are given in the descriptions of Figs. 1 and 2. The analysis of the created diagrams shows a considerable decrease in the mean weekly number of steps (Fig. 1) and weekly number of METs (Fig. 2) after 9 March 2020, which is undoubtedly related to the implementation of the Law on Special Solutions Connected with Prevention, Counteracting and Fighting COVID-19 becoming

106

R. Michnik et al.

legally binding on 8 March 2020. Moreover, on 12 March 2020 the Polish government made a decision on the closure of educational institutions (schools, universities, etc.). The imposed restrictions and widespread panic contributed to the fact that after 9 March, for the following 3 weeks, a very big decline in physical activity was observed in the tested group. That state did not return to the pre-pandemic level for a period of 10 weeks in spite of gradual loosening of restrictions since 20 April 2020. The performed statistical tests ANOVA and ANOVA Friedman showed the occurrence of differences in a weekly number of steps, weekly duration of physical activity, a weekly number of kcal and METs between subsequent weeks ( p ≤ 0.01). The conducted post-hoc Conover-Inmann tests showed the differences in a weekly number of steps between the values obtained for particular weeks 1–7 and the values obtained for weeks 8–17, that is exactly between the values obtained in subsequent weeks before the announcement of the COVID-19 epidemic in Poland and throughout the duration of the pandemic ( p ≤ .05). The post-hoc Tuckey HSD tests proved that statistically significant differences in the duration of physical activity occurred between almost each pair of results obtained in weeks 1–7 and in weeks 8–13 ( p ≤ .05). The exception was the lack of differences in the values obtained for the following week pairs: 4 and 8, 7 and 8, 7 and 13. Similar dependences were observed in subsequent analyzed values: a weekly number of kcal and METs. The performed posthoc Conover-Inmann tests proved that statistically significant differences occurred between each pair of the results obtained in weeks 1–7 and the values obtained in weeks 8–14 ( p ≤ .05).

Fig. 1 Weekly number of steps before and during COVID-19 in Poland (A - suspension of stationary classes in educational institutions, including universities, closure of restaurants, sports facilities and cultural centres; B - introduction of restrictions on movement (lockdown); C - complete closure of parks, boulevards and beaches; D lifting the ban on using parks and forests; E - opening of sports fields (limitation to max. 6 people in the field), cultural centres (libraries, museums, etc.), shopping centres, hotel and tourist sector; F - increase in the limits of the number of users of outdoor sports facilities, possibility of organization of sports events in indoor facilities, re-opening of catering establishments and services sector)

Is the Coronavirus Pandemic Going to ‘Kill’ the Physical Activity of Young People?

107

Table 1 Weekly number of steps and duration of physical activity Weekly number of steps [steps]

Duration of physical activity [min]

Mean

Mean

Std

Min

Max

Min

Max

Before the COVID-19 Week 1 epidemic in Poland Week 2

33941

12719

16627 50634

382

Std 118

229

542

33552

12507

15566 53876

387

144

199

566

Week 3

32178

10043

17214 46431

362

121

160

541

Week 4

30604

10118

18303 45299

347

123

189

566

Week 5

46187

16125

22194 79445

463

122

196

651

Week 6

42583

18880

11620 66018

466

152

103

744

Week 7

31482

14985

12610 59930

322

102

135

491

During the COVID-19 Week 8 epidemic in Poland Week 9

16982

10339

3373 32715

203

104

30

339

10873

7997

2171 27547

152

136

18

475

Week 10

11553

8910

3397 34262

140

100

30

336

Week 11

14193

11427

3724 43153

164

120

35

379

Week 12

12276

9895

3366 38073

154

113

32

350

Week 13

18865

9144

4633 39990

207

96

41

335

Week 14

17718

9401

3766 31700

219

129

33

403

Week 15

20076

14069

3858 53542

283

205

34

691

Week 16

22424

14542

3379 47100

315

234

45

693

Week 17

18842

13038

1803 45400

287

217

17

691

Table 2 Weekly number of kcal and weekly number of METs Weekly number of kcal [kcal]

Weekly number of METs [MET-min/week]

Mean

Mean

Std

Min

Max

Std

Min

Max

Before the COVID-19 Week 1 epidemic in Poland Week 2

1517

534 761

2478

1439

506 801

2482

1608

786 605

2861

1484

644 698

2687

Week 3

1392

455 724

2109

1337

494 561

2132

Week 4

1346

493 634

2158

1279

499 661

2166

Week 5

1787

587 889

2744

1665

424 687

2279

Week 6

1857

711 465

3336

1767

646 360

3019

Week 7

1256

474 611

2007

1169

351 472

1719

During the COVID-19 Week 8 epidemic in Poland Week 9

786

395 135

1222

779

415 104

1260

561

474

69

1622

550

507

64

1770

Week 10

517

348 136

1152

503

369 105

1257

Week 11

585

405 170

1281

588

436 123

1397

Week 12

553

392 130

1295

554

411 111

1251

Week 13

746

339 186

1231

715

332 144

1179

Week 14

851

459 151

1397

847

520 117

1482

Week 15

1232

980 155

3335

1185

948 119

3018

Week 16

1357

1140 205

3481

1369

1163 162

3217

Week 17

1233

1048 100

3180

1275

1105

60

3469

108

R. Michnik et al.

Fig. 2 Weekly number of METs before and during COVID-19 in Poland (A - suspension of stationary classes in educational institutions, including universities, closure of restaurants, sports facilities and cultural centres; B - introduction of restrictions on movement (lockdown); C - complete closure of parks, boulevards and beaches; D lifting the ban on using parks and forests; E - opening of sports fields (limitation to max. 6 people in the field), cultural centres (libraries, museums, etc.), shopping centres, hotel and tourist sector; F - increase in the limits of the number of users of outdoor sports facilities, possibility of organization of sports events in indoor facilities, re-opening of catering establishments and services sector)

4 Discussion Physical Activity During Coronavirus Pandemic SARS CoV-2 The outbreak of the coronavirus pandemic and resulting social distancing led to a dramatic drop in physical activity, what indicate data presented in the Community Mobility Report prepared by Google [22]. Reduced physical activity was observed in the investigations presented in this article. The first 7 weeks, i.e. the period preceding the pandemic, witnessed the average weekly number of steps at a level of approx. 30604–46187, which translates into around 4372–6598 steps per day. According to the Tudor-Locke et al. [14], the above-mentioned data classify the subjects as persons having the sedentary lifestyle. The difference between the minimum number and the maximum number oscillated within the range of: 26996 (week 4) even up to 57251 (week 5), which resulted from a different level of activity of the tested persons. The announcement of the state of epidemic and the introduction of restrictions affecting people’s lifestyles led to a clear decrease in physical activity - both with reference to a number of steps and the duration of physical activity. In the 8th week of measurements, Poland introduced first lockdown restrictions due to the widespread epidemic, including the suspension of stationary educational classes at universities, closure of restaurants, sports facilities (gyms, swimming pools, fitness clubs and dance clubs) and cultural institutions (cinemas, museums, libraries) (Fig. 1 - item A), which resulted in a 53%

Is the Coronavirus Pandemic Going to ‘Kill’ the Physical Activity of Young People?

109

reduction of a weekly number of steps from the average value of 35790 ± 13625 down to 16982 ± 10339 as well as a 48% decrease in weekly duration of physical activity from the average value of 390 ± 126 min down to 203 ± 104 min. In the first week of the pandemic, the total activity was at a level of 779 ± 415 MET-min/week, which in comparison with the previous state (1169–1767 METmin/week) caused 1.5–2.3-fold reduction in a weekly value of MET. Subsequent restrictions (Figs. 1 and 2), i.e.: B - restrictions on mobility (lockdown) and C - categorical closure of parks, boulevards and beaches during the 9th, 10th, 11th and 12th measurement week contributed to a considerable drop in physical activity. The applied restrictions led to the reduction of a weekly number of steps even up to 10873 steps, which on average gave daily only around 1550 steps - compared to the period preceding the epidemic, it is a number 2.8–4.3-times smaller. The limitations on doing sports in the open-air contributed also to a substantial decline in the duration of physical activities. The period of the above-mentioned 4 weeks witnessed the average duration of physical activities amounting to 153 ± 117 min, i.e. only approx. 2.5 h per week. Summing up the time of walking and additional physical activities, the metabolic equivalent of task equalled 549 ± 431 MET, which signifies over a 3-fold drop in the energy expenditure. The mean value of burnt calories amounted only to 554 ± 404 kcal. The loosening of the lockdown restrictions consisting in the lifting of the ban on the entry to parks and forests in the 13th week (Figs. 1–2 Item D), i.e. the return of the possibility of doing physical activities in the open air, led to a 50% increase in a weekly number of steps and prolongation of physical activities by 54 min in comparison with the physical activities in the first weeks of the epidemic. Simultaneously, the total energy expenditure was increased by 166MET-min/week and the number of burnt calories by 192 kcal. Further lifting of restrictions enabling: E - the use of sports fields (limitation up to max. 6 persons in the field), cultural centres (libraries, museums, etc.), shopping centres, hotel and tourist sector, and also F - an increase in the limited number of users of outdoor sports facilities, organization of sports events in indoor sports facilities and re-opening of catering establishments (such as restaurants, cafés, pubs) and service establishments (like barber’s and hairdresser’s, beauty parlours, etc.) - caused the rise in energy expenditure up to a value of 1369 MET-min/week (week 16), i.e. almost two-fold in relation to the data obtained in the first week of the epidemic (Fig. 2). Despite a considerable hike in the physical activity in weeks 14–17 by increasing the number of steps up to 22 thousand and the duration of physical activity up to 315 min per week, the situation did not fully return to the pre-epidemic state of physical activity in Poland. Consequences of the Sedentary Lifestyle The review of the data which were made available by Google [22], Fitbit [19], and also the results of the current investigations indicate that the restrictions introduced during the SARS CoV-2 pandemic had impact on a considerable decline in the physical activity of the society. Such a dramatic fall in physical activities, which lasts for a long period of time, may have serious consequences for health. Undoubtedly, lower physical activity, mainly a smaller number of force and resistance exercises, may negatively influence the mineral density of bones. Numerous

110

R. Michnik et al.

scientific studies show that higher mineral density of bones as well as a smaller risk and lower frequency of fractures occur in physically active persons [4]. Moreover, physical exercises are vital for maintaining proper mass of muscles through activation of the synthesis of muscle proteins [2]. Olsen et al. [12], Knudsen et al. [7], Krogh-Madsen et al. [8] point out to the fact that a decrease or rapid decrease in the number of steps down to 750–1500 steps per day causes a loss of muscle mass in lower limbs in healthy young adults and the elderly within two weeks. In the conducted studies, week 9 was reported as the week with the average daily number of steps at a level of approx. 1500 for the whole group. In the following week, the daily number of steps maintained a similar level (on average 1650 steps). It should also be noted that almost half of the group did not exceed the number of 1000 steps in two subsequent weeks. It can be suspected that the subjects suffered from the loss of muscle mass in their lower limbs. Moreover, Krogh-Madsen et al. [8] proved that the reduction of daily physical activity to a similar level to the above-described one but with the maintenance of energy balance may lead to a decrease in sensitivity to insulin, decrease in VO2 maximum by 7 mL/kg/min and reduction of the fat-free mass of lower limbs by 0.5 kg in healthy young male adults [8]. Olsen et al. also showed that after 21 days of reduced physical activity, the subjects showed reduced sensitivity to insulin, weakening of the postprandial metabolism of lipids and increased mass of intra-abdominal fat [12]. A positive energy balance in the period of physical inactivity had a considerable impact on metabolic regulation, body content, efficiency of muscles and the profile of cardio-metabolic risk. The combination of the positive energy balance with the lack of physical activity leads to the accumulation of fat in visceral septa [12] and the catabolism of fat-free body mass [9]. Excessive consumption of food, lack of activity and fat accumulation result in a low inflammatory reaction and increasing oxidative stress [10]. Biolo et al. [1] proved that during 35 days of rest in bed, in healthy young subjects, at different levels of energy consumption, the growth of fat tissue is connected with the largest loss of the mass of skeletal muscles. Many studies analyzed the impact of the reduction of steps on health parameters proving that even a short-term decrease in physical activity has a negative influence on the metabolism of skeletal muscles proteins and carbohydrates [2, 11]. Staying in a confined space during isolation causes psychological stress which leads to endocrinological changes. Such changes influence the body’s immunity and have other negative psychological effects, including the symptoms of post-traumatic stress, confusion and anger. Limitations of this Work and Directions of Further Research The authors fully realise the limitations of this work. This work presents in a longitudinal way (17 weeks) the investigations of physical activity of 11 students before and during the pandemic of coronavirus SARS CoV-2. The limitations include a small number of tested persons. Another disadvantage may be the recording of the number of steps by means of two types of applications (Samsung Health and Health for the IOS system). The presented results could be supplemented with some studies on the

Is the Coronavirus Pandemic Going to ‘Kill’ the Physical Activity of Young People?

111

impact of the SARS CoV-2 coronavirus pandemic on mental health of both young and elderly people.

5 Conclusions On the basis of the analysis of the obtained results and literature studies, the following final conclusions were formulated: • physical activity of the tested subjects before epidemic COVID-19 was at a low level. A weekly number of steps amounted to 30604–46187, which gives around 4372-6598 steps per day. • Epidemic COVID-19 led to a ‘dangerous’ drop in physical activity in the tested persons. The applied restrictions led to the reduction of a weekly number of steps even up to 10873 steps, which on average gave daily only around 1550 steps. In almost half of the subjects this value did not exceed 1000 steps per day in two subsequent weeks. • During the epidemic, the researchers observed a considerable decrease in the duration of additional physical activities, which translated into the reduction in energy expenditure. • Physical activity of the subjects did not return to the pre-epidemic level for a period of 10 weeks despite loosening of the restrictions. • Reduction in physical activity to the above-described level may have a negative impact both on physical and mental health.

References 1. Biolo G et al (2008) Positive energy balance is associated with accelerated muscle atrophy and increased erythrocyte glutathione turnover during 5 wk of bed rest. Am J Clin Nutr 88:950–958 2. Bowden Davies KA et al (2019) Reduced physical activity in young and older adults: metabolic and musculoskeletal implications. Ther Adv Endocrinol Metab 10:2042018819888824 3. Brooks SK et al (2020) The psychological impact of quarantine and how to reduce it: rapid review of the evidence. Lancet 395:912–920 4. Gregg EW, Pereira MA, Caspersen CJ (2000) Physical activity, falls, and fractures among older adults: a review of the epidemiologic evidence. J Am Geriatr Soc 48:883–893 5. Jochymczyk-Wo´zniak K, Nowakowska K, Michnik R, Gzik M, Kowalczykowski D (2019) Three-dimensional adults gait pattern–reference data for healthy adults aged between 20 and 24. In: Tkacz E, Gzik M, Paszenda Z, Pie˛tka E (eds) IBE 2018, vol 925. AISC. Springer, Cham, pp 169–176. https://doi.org/10.1007/978-3-030-15472-1_19 6. Jordan JL, Holden MA, Mason EE, Foster NE (2010) Interventions to improve adherence to exercise for chronic musculoskeletal pain in adults. Cochrane Database Syst Rev. 1 7. Knudsen SH, Hansen LS, Pedersen M et al (2012) Changes in insulin sensitivity precede changes in body composition during 14 days of step reduction combined with overfeeding in healthy young men. J Appl Physiol 113:7–15

112

R. Michnik et al.

8. Krogh-Madsen R et al (2010) A 2-wk reduction of ambulatory activity attenuates peripheral insulin sensitivity. J Appl Physiol 108:1034–1040 9. Lovejoy JC et al (1999) Low-dose T3 improves the bed rest model of simulated weightlessness in men and women. Am J Physiol Endocrinol Metab 277(2):E370–E379 10. Narici M, De Vito G, Franchi M et al (2020) Impact of sedentarism due to the COVID-19 home confinement on neuromuscular, cardiovascular and metabolic health: physiological and pathophysiological implications and recommendations for physical and nutritional countermeasures. Eur J Sport Sci 21(4):614–635 11. Oikawa SY, Holloway TM, Phillips SM (2019) The impact of step reduction on muscle health in aging: protein and exercise as countermeasures. Front Nutr 6:75 12. Olsen RH, Krogh-Madsen R, Thomsen C, Booth FW, Pedersen BK (2008) Metabolic responses to reduced daily steps in healthy nonexercising men. JAMA, J Am Med Assoc 299:261–1263 13. Schnohr P, Grønbæk M, Petersen L, ia Sørensen T, Ole Hein H (2005) Physical activity in leisure-time and risk of cancer: 14-year follow-up of 28,000 Danish men and women. Scand J Public Health 33:244–249 14. Tudor-Locke C, Craig CL, Thyfault JP, Spence JC (2013) A step-defined sedentary lifestyle index: 0.05). Table 2 shows the determined values of the total length of the spine and the length of its individual sections - the thoracic and lumbar sections. Statistical analysis

132

H. Zado´n et al.

Table 1 Mean values of the angle of total trunk inclination, thoracic kyphosis, lumbar lordosis and the angle of the pelvis (pelvic inclination angle - designations: (–) anteversion, (+) retroposture) while standing and sitting in 3 different positions, * - normal distribution Standing position I. Slumped sitting II. Preferred III. Upright position sitting position sitting position Mean ± SD Total trunk inclination [deg] Pelvic inclination [deg] Thoracic kyphosis [deg] Lumbar lordosis [deg]

2.3 ± 1.9

16.7 ± 7.5*

7.0 ± 3.6*

4.4 ± 3.6

–16.1 ± 5.0*

5.5 ± 9.1

4.8 ± 8.1*

–1.1 ± 6.5*

24.9 ± 11.4*

47.0 ± 20.0

34.8 ± 11.9

26.0 ± 11.1

24.5 ± 9.8*

0.9 ± 2.2

4.3 ± 6.8

9.5 ± 11.4

Table 2 Mean values of the total length of the spine, thoracic and lumbar spine section. * - normal distribution Standing position I. Slumped sitting II. Preferred III. Upright position sitting position sitting position Mean ± SD Total spine length 497 ± 41* [mm] Total thoracic 371 ± 35* length [mm] Total lumbar 93 ± 20* length [mm]

577 ± 46*

543 ± 43*

525 ± 42*

395 ± 43*

371 ± 38*

364 ± 34*

133 ± 26*

126 ± 22*

118 ± 19*

showed that for all parameters presented in the table for each position, there is a normal distribution. In the tests performed, in the case of the total length of the spine and the total length of the thoracic segment, statistically significant differences were not observed between the standing position and the III position - upright, and between the II position - preferred position and the III position - upright position with the adopted significance level α = 0.05. For the parameter total length of the lumbar spine, there were differences between the standing position and all sitting positions (p < 0.05), but they were not found between the I - slumped and the II preferred position (p = 0.56), and between the II - preferred and III - upright positions (p = 0.53). For most of the examined parameters, smaller differences were recorded between position II and position III than between position II and position I.

Body Postures During Sitting in Different Positions

133

4 Discussion During development, the spine, due to the effect of gravity on the body, is displaced in the sagittal plane, shaping its physiological curves - cervical lordosis, thoracic kyphosis, and lumbar lordosis. These curves are necessary for the proper functioning of both the muscular and skeletal systems. The correct shape of the curves allows you to maintain the vertical position of the body, the correct position and the weight of the head, upper limbs, shoulder girdle, pelvis, lower limbs and feet [23]. Because during static standing and sitting, the curvature changes are associated with higher mechanical loading, neutral spinal alignment is regarded as the optimal loading [4]. In this study, it was noticed that the manner of adopting a sitting position significantly influences the change in the curvature of the spine, which, especially in people practicing a sedentary lifestyle, may adversely affect the change of body posture and play a significant role in the development and maintenance of pain in the spine [15, 16]. The kinematics tests carried out have shown that the adoption of a sitting position may lead to an almost 2-fold increase in thoracic kyphosis compared to a standing position. The highest angle of kyphosis was recorded during the assumption of position I - slumped (47.0 ± 20.0◦ ), then this angle decreased with the adoption of position II - the preferred (34.8 ± 11.9◦ ) until it was reached during position III upright thoracic kyphosis, amounting to 26.0 ± 11.1◦ . It should be noted that the value of thoracic kyphosis in standing position is similar to kyphosis in sitting position III upright but not to kyphosis to position II - preferred, which shows that in the studied group of people during daily sitting, thoracic kyphosis grew. Literature data indicate that the increase in thoracic kyphosis, which is currently mainly observed in the elderly population, affects to increased incidence of intrascapular pain [8], increased body sway, gait unsteadiness, and a higher risk of falls [11, 22]. In contrast, in the case of lumbar lordosis, it flattens with the adoption of a sitting position. Sitting reduces lordosis from 9.5 ± 11.4◦ for position III - upright to 0.9 ± 2.2◦ for the position I slumped, while the mean angle of lumbar lordosis when standing was 24.5 ± 9.8◦ . A 2–3-fold reduction in the lordosis angle value while sitting compared to standing position was also presented in the work of Hey et al. [10] and de Carvalho et al. [2]. The large values of deviations from the mean recorded in this study indicate different interpretations of individual sitting positions by the test participants. The shape of the lumbar lordosis is undoubtedly influenced by the position of the pelvis. Literature data confirm that setting the pelvis in a retroverted position leads to the loss of the preferred curvature of the lumbar section, i.e. it is the reason for the reduction of the lordosis angle [2, 4, 7, 9, 19]. In these studies, it was noticed that the adoption of a sitting position causes the pelvis to be retroverted in the case of position I - slumped (5.5 ± 9.1◦ ) and position II - the preferred (4.8 ± 8.1◦ ) or a slight anterversion in the case of position III - upright (−1.1 ± 6.5◦ ). Michnik et al. [19] indicate that despite the positive spine straightening effect that occurs during pelvic retroversion, attention should be paid to the generation of high values of the resultant reaction forces in the intervertebral joints. According to the results of computer simulations presented by the authors, it can be assumed that during the

134

H. Zado´n et al.

sitting position in which an increase in the pelvic retroposition has been noticed, the loads occurring in the lumbar spine may increase in relation to the standing position. However, additionally, by analyzing individual sitting positions, it can be concluded that the position I - slumped due to the greatest pelvic retroposition will be the cause of the greatest resultant values of reaction forces in the lumbar region. The influence of slumped posture while sitting on the increase of loads was also recorded in experimental studies conducted by Wilke et al. [26]. Wilke et al. indicated that adopting a slumped sitting position may lead to a 3-fold increase in load compared to a relaxed sitting position [26]. The trunk inclination angle can also affect the load values in the lumbar spine. Nowakowska et al. [20] showed that with increasing trunk inclination, the load on the lumbar region increases. The assumption of the sitting position resulted in an almost 2–7-fold increase in the trunk inclination compared to the standing position,according to Nowakowska et al.[20], it may be another reason for increasing spinal loads during sitting. The trunk inclination angle increased with the transition from the sitting position III - upright (4.4 ± 3.6◦ ) through the position II - preferred (7.0 ± 3.6◦ ) to the position I - slumped, where the average angle the inclination was as high as 16.7 ± 7.5◦ . Differences between sitting positions were also noticed in the total length of the spine and the lengths of its individual sections. These changes may result from changes in body posture as well as from compression and creep of the intervertebral disc [1, 13]. Literature studies show that the amount of loss/increase in the length of the spine and its sections are proportional to the increase or decrease in compressive loads acting on the spine [13]. In this study, it was observed that the longest mean total length of the spine (577 ± 46 mm), the thoracic spine (395 ± 43 mm) and the lumbar spine (133 ± 26 mm) was recorded when taking the position I - slumped, while the lowest during position III - upright where the mean total length of the spine was: 525 ± 42 mm, the thoracic spine: 364 ± 34 mm and the lumbar spine: 118 ± 19 mm. The total length of the spine was extended on average by 16 ± 4% when taking the position I - slumped, by 9 ± 4% in position II - the preferred, and by 6 ± 4% during position III - upright compared to the total length of the spine recorded while standing (100%). On the other hand, when comparing the lengths of individual sections of the spine, it was noticed that while sitting, the length of the lumbar section increased relative to the standing position, which may be related to changes in the position of the pelvis. The impact of decreasing spine lengths and segments of the spine while sitting was also observed in experimental studies by Kourtis et al. [13] and Althoff et al. [1]. Smaller differences in the parameters: total trunk inclination, thoracic kyphosis, and total spine length were observed between position II and position III than between position II and position I, suggesting that the subjects assumed the preferred position closer to the upright position than to the slumped position.

Body Postures During Sitting in Different Positions

135

5 Conclusions As part of this study, a total of 36 people’s body postures were examined in three different positions while sitting: the preferred position, slumped position, and upright position. The conducted comparative analysis allowed for the formulation of the following conclusions: – the sitting position causes the position of the pelvis to be changed: in retroposition (position I - slumped: 5.5 ± 9.1◦ and position II - preferred: 4.8 ± 8.1◦ ) or slight anteversion (position III - upright: –1.1 ± 6.5◦ ), which reduces the angle of lumbar lordosis in relation to the standing position, – assuming a sitting position resulted in almost 2–7-fold increase in trunk inclination compared to the standing position, which may increase the load on the lumbar spine, – during sitting the spine lengthens in relation to a standing position, the total length of the spine decreases with the transition from the position I - slumped through the preferred position - II to position III - upright. The presented research work has some limitations. During the tests, the manner of sitting was not imposed - subjects were tasked with adopting a slumped, neutral and upright posture, which could cause different interpretations of the task. The results were analyzed only in specific age groups (20–26 years) and taking into account basic kinematic data. In the further study, it would be also necessary to investigate the effects of body mass, body height and age on the size of the pelvic inclination angle and the size of the curvatures of the spine. Acknowledgements Publication supported by Own Scholarship Fund of the Silesian University of Technology in the year 2019/2020, grant number: 32/FSW18/ 0003-03/2019.

References 1. Althoff I, Brinckmann P, Frobin W, Sandover J, Burton K (1992) An improved method of stature measurement for quantitative determination of spinal loading. Spine 17(6):682–693 2. De Carvalho DE, Soave D, Ross K, Callaghan JP (2010) Lumbar spine and pelvic posture between standing and sitting: a radiologic investigation including reliability and repeatability of the lumbar lordosis measure. J Manipulative Physiol Ther 33(1):48–55 3. Chowanska J, Kotwicki T, Rosadzinski K, Sliwinski Z (2012) School screening for scoliosis: can surface topography replace examination with scoliometer? Scoliosis 7(1):1–7 4. Claeys K, Brumagne S, Deklerck J, Vanderhaeghen J, Dankaerts W (2016) Sagittal evaluation of usual standing and sitting spinal posture. J Bodyw Mov Ther 20(2):326–333 5. Drygas W, Kwa´sniewska M, Szcze´sniewska D, Kozakiewicz K, Głuszek J, et al (2015) Ocena poziomu aktywno´sci fizycznej dorosłej populacji Polski. Wyniki programu WOBASZ. Polish Heart J. 63(6):636–640 6. Drza-Grabiec J, Truszczy´nska A, Fabja´nska M, Trzaskoma Z (2016) Changes of the body posture parameters in the standing versus relaxed sitting and corrected sitting position. J Back Musculoskel Rehabil 29(2):211–217

136

H. Zado´n et al.

7. Endo K, Suzuki H, Nishimura H, Tanaka H, Shishido T, Yamamoto K (2012) Sagittal lumbar and pelvic alignment in the standing and sitting positions. J Orthop Sci 17(6):682–686 8. Griegel-Morris P, Larson K, Mueller-Klaus K, Oatis CA (1992) Incidence of common postural abnormalities in the cervical, shoulder, and thoracic regions and their association with pain in two age groups of healthy subjects. Phys Therapy 72:425–431 9. Hayden AM, Hayes AM, Brechbuhler JL, Israel H, Place HM (2018) The effect of pelvic motion on spinopelvic parameters. Spine J 18(1):173–178 10. Hey HWD, Teo AQA, Tan KA, Ng LWN, Lau LL, Liu KPG, Wong HK (2017) How the spine differs in standing and in sitting-important considerations for correction of spinal deformity. Spine J 17(6):799–806 11. Hirose D, Ishida K, Nagano Y, Takahashi T, Yamamoto H (2004) Posture of the trunk in the sagittal plane is associated with gait in community-dwelling elderly population. Clin Biomech 19(1):57–63 12. Korakakis V, O’Sullivan K, O’Sullivan PB, Evagelinou V, Sotiralis Y, Sideris A, Sakellariou K, Karanasios S, Giakas G (2019) Physiotherapist perceptions of optimal sitting and standing posture. Musculoskel Sci Pract 39:24–31 13. Kourtis D, Magnusson ML, Smith F, Hadjipavlou A, Pope MH (2004) Spine height and disc height changes as the effect of hyperextension using stadiometry and MRI. Iowa Orthop J 24:65–71 14. Kowaiski IM, Protasiewicz-Faldowskaa H, Siwil P, Zaborowska-Sapetaa K, Dqbrowsk A, Kiuszczyñski M, Raistenskis J (2013) Analysis of the sagittal plane in standing and sitting position in girls with left lumbar idiopathic scoliosis. Polish Ann Med 20(1):30–34 15. Levine JA (2015) Sick of sitting. Diabetologia 58(8):1751–1758 16. Lis AM, Black KM, Korn H, Nordin M (2007) Association between sitting and occupational LBP. Eur Spine J 16(2):283–298 17. Loyen A, Van Der Ploeg HP, Bauman A, Brug, J, Lakerveld J (2016) European sitting championship: prevalence and correlates of self-reported sitting time in the 28 European Union Member States. PLoS One 11(3):e0149320 18. Matyja M, Saulicz E, Saulicz M, Kokosz M, Gnat R, Kuszewski M, Bereszko J, Gogola A (2010) An assessment of rotational mobility of the trunk among teenagers with faulty posture. J Hum Kinet 24(1):43–49 19. Michnik R, Zado´n H, Nowakowska-Lipiec K, Jochmyczyk-Wo´zniak K, My´sliwiec A, Mitas AW (2020) The effect of the pelvis position in the sagittal plane on loads in the human musculoskeletal system. Acta Bioeng Biomech 22(3):33–42 20. Nowakowska K, Gzik M, Michnik R, My´sliwiec A, Jurkoj´c J, Sucho´n S, Burkacki M (2017) The loads acting on lumbar spine during sitting down and standing up. In: Advances in intelligent systems and computing. Springer, Heidelberg, pp 169–176 21. O’Sullivan K, O’Sullivan P, O’Sullivan L, Dankaerts W (2012) What do physiotherapists consider to be the best sitting spinal posture? Man Ther 17(5):432–437 22. Sinaki M, Brey RH, Hughes CA, Larson DR, Kaufman KR (2005) Balance disorder and increased risk of falls in osteoporosis and kyphosis: significance of kyphotic posture and muscle strength. Osteoporos Int 16(8):1004–1010 23. Tokpinar A, Ülger H, Yilmaz S, Acer N, Ertekin T, Görkem SB, Güler H (2019) Examination of inclinations of the spine at childhood and adolescence. Folia Morphol 78(1):47–53 24. Varo JJ, Martínez-González MA, de Irala-Estévez J, Kearney J, Gibney M, Martínez JA (2003) Distribution and determinants of sedentary lifestyle in the European Union. Int J Epidemiol 32(1):138–146 25. Watanabe S, Kobara K, Yoshimura Y, Osaka H, Ishida H (2014) Influence of trunk muscle co-contraction on spinal curvature during sitting. J Back Musculoskel Rehabil 27(1):55–61 26. Wilke HJ, Neef P, Caimi M, Hoogland T, Claes LE (1999) New in vivo measurements of pressures in the intervertebral disc in daily life. Spine 24(8):755–762 27. World Health Organization (2009) Global health risk: mortality and burden of disease attributable to selected major risks. World Health Organization, Geneva

Engineering of Biomaterials

Engineering of biomaterials is a key field of science during all design works, both for diagnostic and therapeutic devices as well as for implants. This is evidenced by numerous works presented in scientific journals, conference proceedings and monographs synthesizing knowledge. The chapter “Engineering of Biomaterials” in this monograph presents the results of scientific and research works concerning both the structure and functional properties of various biomaterials and the possibilities of their use in selected areas of medicine. Mechanical properties of biomaterials, creation of new photopolymers or development of implant production technology are examples of the subject matter of the included articles. Furthermore, issues regarding the possibility of using 3D printing technology and its influence on bone phantom features were also presented.

Adhesion of Staphylococcus Aureus on Various Biomaterial Surfaces Marcin Basiaga, Zbigniew Paszenda, Marcin Kaczmarek, Witold Walke, Agata Sambok-Kiełbowicz, Wojciech Kajzer, Anna Taratuta, Julia Lison, ´ Magdalena Szindler, and Alicja Kazek-Ke˛sik

Abstract At present, it is believed that 60 to 80% of the infections that humans encounter after implantation are related to the formation of biofilms. Biofilm-forming bacteria are particularly resistant to antibiotics and human immune mechanisms. This process is multistage, conditioned by the properties of the microorganisms that make it up, and by the structure and properties of the colonized materials. Therefore, it was proposed to apply the Z n O antibacterial layer on the substrate made of T i 6 Al7 N b alloy by the ALD method. To evaluate the proposed surface modification, studies of electrochemical properties were carried out. In addition, biological studies were performed with the reference bacteria Staphylococcus aureus (ATCC 25923). The results show that the number of bacterial colonies adhered to the tested surfaces depends on the variant of used surface modification. Keywords ALD method · Ti alloys · Sandblasted · Electropolished · S. aureus

1 Introduction The biofilm formation begins with the adhesion of bacteria to the implant surface. The basic strategy against the development of the surrounding inflammation is to prevent the adhesion of microorganisms [4]. This process can be divided into two phases—reversible and irreversible. In the first phase, the bacteria adhere to the surface of the implant and start to secrete protective substances, and in the second phase, after good encapsulation, the biofilm grows. The microorganisms are attached to dense and slimy barriers composed of sugars and proteins. Biofilm is a separating M. Basiaga (B) · Z. Paszenda · M. Kaczmarek · W. Walke · A. Sambok-Kiełbowicz · W. Kajzer · A. Taratuta · J. Liso´n Faculty of Biomedical Engineering, Silesian University of Technology, Zabrze, Poland e-mail: [email protected] M. Szindler Faculty of Mechanical Engineering, Silesian University of Technology, Gliwice, Poland A. Kazek-Ke˛sik Faculty of Chemistry, Silesian University of Technology, Gliwice, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_15

139

140

M. Basiaga et al.

of bacteria from external factors. It allows for survival despite antibiotics or organism defense mechanisms. Limiting the adhesion of bacteria to the implant surface is crucial because bacteria that cannot adhere quickly to the surface are quickly killed by lymphocytes [5]. The creation of biofilms can often lead to disturbances of wound healing, inflammatory diseases, periodontal disease, cystic fibrosis, chronic acne, osteoarthritis, or osteomyelitis [2, 6]. About 60–70% of the microorganisms which are attached to the implant surface have connection with current nosocomial infections [6]. At the same time, most infections are connected with creating biofilm on biomaterial surfaces. Staphylococcus are popular microorganisms whose multiplication leads to prosthetic joint infections and heart implants. These kinds of bacteria are responsible for about 40–50% infections in the cardiovascular system and 50–70% infections on catheters [2, 5, 7]. Titanium and titanium alloys, as materials used for the production of implants, are corrosion-resistant, non-allergenic, and nontoxic, when they do not contain vanadium. Lots of modification of its surface increases its abrasion resistance and allows it to avoid allergenic effects [8]. Additionally, its good biocompatibility relies on a thin layer of titanium oxide on its surface. However, a biofilm will begin to form on each implant placed in the human body, regardless of the type of material, within a few minutes of implantation. Susceptibility to adhesion of bacteria or even human cells depends primarily on the surface structure and composition of the biomaterial. In [1] the authors examined titanium implants modification. As a result, the investigation showed that the presence of Si O2 − T i O2 layer on the sample surface caused a decreasing ratio of S. epidermidis adhesion. T i N layer increases abrasion resistance but does not prevent bacterial adhesion. Zhang et al. [10] proved that T i − 3Cu has great anti corrosion properties due to its strong anti adhesive properties against S. aureus, thanks to the formation of T i 2 Cu build-up. In [9] T i O2 , T i O2 with zinc, and Z n O coatings have been placed on the T i implant surface using a cathodic arc deposition system. Then osteoblasts and S. aureus were grown on each surface. Antibacterial tests indicated no visible difference between the uncoated T i plate and T i O2 coating, but with Z n O and T i(Z n)O2 coatings a much smaller number of bacteria was observed. Biocompatibility studies showed that T i O2 and T i(Z n)O2 coatings increase cell viability and proliferation compared to uncoated T i plate and Z n O coating. However, significant inhibition of cell viability was found in the case of Z n O coating, which may result in cytotoxicity of excessive zinc concentration. This leads to the conclusion that the T i(Z n)O2 coatings with a lower Z n content (7.6 ± 1.3%) enhances antibacterial properties while maintaining good biocompatibility with osteoblasts. All these studies confirm the necessity and effectiveness of titanium surface modification for implant applications. In this paper, it was decided to apply the antibacterial layer Z n O on a substrate made of T i 6 Al7 N b alloy by ALD method. To validate the obtained results, electrochemical properties and biological tests with standard bacteria Staphylococcus aureus were proposed.

Adhesion of Staphylococcus Aureus on Various Biomaterial Surfaces

141

2 Materials and Methods Titanium alloy (T i 6 Al7 N b) samples obtained from a bar of diameter d = 14 mm were used in the study. The samples were sandblasted and electropolished. The sandblasting process was carried out using glass balls with diameters ranging from 50 to 150 µm on a Micra 2 precision sandblaster. The applied time was equal to t = 2 min and the pressure was p = 3 bar. Electrochemical polishing was performed by using a solution of phosphate sulfate. A Z n O layer was applied on the prepared substrates by the ALD method. Zinc oxide thin films were deposited by the atomic layer deposition method using a R-200 standard system from Picosun. The chamber temperatures were 100 and 300 ◦ C (layer thickness for 500 cycles about 100 nm and 1500 cycles about 250 nm).

2.1 Potentiodynamic Test The tests were carried out as recommended by the ASTM F2129 standard. The test set-up consisted of the VoltaLab PGP201 potentiostat, the reference electrode (KP113 saturated calomel electrode SCE), the auxiliary electrode (PtP-201 platinum electrode), the working electrode (test sample) and a PC with VoltaMaster 4 software. The corrosion tests started with establishing the open circuit potential E ocp at currentless conditions during the time T = 120 min. The polarization curves were recorded starting with the initial potential value, E init = E ocp − 100 mV. The potential changed along the anode direction at the rate of 0.167 mV/s. The studies were carried out in the range of potential changes from E init to +4000 mV for each sample. On the basis of the obtained curves, the corrosion potential E corr was determined, and the value of the polarization resistance R p was calculated with the use of the Stern method.

2.2 Impedance Test To obtain more information about the electrochemical properties of the analyzed surfaces, electrochemical impedance spectroscopy (EIS) studies were carried out. The measurement set-up consisted of the Auto Lab PGSTAT 302N system equipped with FRA2 (frequency response analyser) module, the reference electrode (KP-113 saturated calomel electrode SCE), the auxiliary electrode (PtP-201 platinum electrode), and the working electrode—test sample. The open circuit potential E ocp was determined for each sample in 120 min. The EIS studies ensured information on

142

M. Basiaga et al.

the surface layer tightness and its barrier properties, as related to the Ringer solution. The impedance spectra were presented as Nyquist plots for different frequencies (104 –10−3 Hz) and as Bode diagrams. The sinusoid voltage amplitude of the activating signal was 10 mV. The obtained EIS spectra were interpreted after fitting to the electrical equivalent circuits by means of the least squares method.Both the potentiodynamic and impedance tests were carried out in the Ringer solution (N aCl – 8.6 g/l, K Cl – 0.3 g/l, CaCl2 · 6H2 O – 0.48 g/l), (250 g/l), supplied by Baxter, at T = 37 ± 1 ◦ C and p H = 7 ± 0.2.

2.3 Testing the Adhesion of Microorganisms to the Surface Studies were performed using the standard bacterium Staphylococcus aureus (ATCC 25923). Bacteria in the amount of ∼5 · 106 CFU/ml in a liquid medium (Graso, Biotech, Poland) were applied to the tested material and incubated in an incubator (240CLW, POL-EKO) at 37 ◦ C for 4 h. Then, the culture medium was gently withdrawn, the samples were rinsed with sterile water and transferred to a new culture plate. The samples were then vigorously mixed with the trypsin-containing solution. The next step involved taking 100 µl of the test solution with microorganisms and preparing a series of dilutions in a sterile 0.9% NaCl solution. 100 µl of the solution was seeded on Müller-Hinton microbial media (Agar, Diag-Med) and placed in an incubator at 37 ◦ C for 18 h, then bacterial colonies were counted. TCPS was used for the culture control and reference adhesion of S. aureust ATCC 25923 bacteria.

3 Results and Discussion 3.1 Potentiodynamic Test The results of potentiodynamic tests of pitting corrosion resistance are presented in Table 1 and in Fig. 1. Based on the obtained results, a differentiated pitting corrosion resistance of samples with the Z n O layer was found depending on the applied temperature of application. It was found that for each variant of samples with a Z n O layer, an improvement in corrosion resistance was observed compared to samples without the layer. Moreover, it was observed that with the increase of the application temperature and the number of cycles, the resistance to pitting corrosion would increase.

Adhesion of Staphylococcus Aureus on Various Biomaterial Surfaces Table 1 Results of potentiodynamic tests—mean values Lp. Surface modification E corr , mV Temp, ◦ C Cycles Sandblasted 1 Initial state 2 100 4 8 300 10 Electrochemical polished 1 Initial state 2 100 4 8 300 10

143

E tr , mV

R p , k · cm2

250 335 740 6030 8222

500 1500 500 1500

–136 –88 –190 –52 636

– 1733 – – –

500 1500 500 1500

54 –180 –179 –152 109

– – – – –

91 115 332 103 319

Fig. 1 Polarization curves regarding T i 6 Al7 N b after sandblasting at initial condition and after Z n O ALD surface modification

3.2 Impedance Test The results of the impedance tests are presented in Table 2 and Figs. 2, 3 and 4. In the case of the impedance tests, a similar trend was observed as in the case of the potentiodynamic tests. It was found that the value of the ion transition resistance is greater with increasing temperature and the number of cycles. Model a: The passive film and electrical double layer are assumed to be shown with ideal capacitive behavior or non-ideal capacitive behavior. R por e and C P E por e are representatives of the electrical porous layer where as Rct and C P E dl represent the resistive and nonideal capacitive behavior of the passive film (double layer). According to this model, the two R, C P E elements are connected in series with the

144

M. Basiaga et al.

Table 2 Results of EIS Sample

Rs , k· cm2

R por e , k· cm2

C P E por e

E OC P , mV

Rct , C P E dl k · cm 2

Y0 , −1 cm−2 s−n

n

Y0 , −1 cm−2 s−n

n

Sandblasted

68

5

0.1128E–5

0.69

262

0.6062E–4

0.77

100–500

67

54

0.2587E–4

0.71

452

0.4902E–5

0.81

–235 –167

100–1500

69

61

0.2001E–4

0.58

840

0.1064E–5

0.75

–80

300–500

68

85

0.1097E–4

0.77

383

0.1964E–4

0.79

–83

300–1500

67

36

0.4037E–4

0.67

415

0.2458E–5

0.84

–127

Electrochemical polished

66

2

0.5478E–5

0.71

294

0.3658E–4

0.77

–99

100–500

69

46

0.8790E–5

0.70

394

0.3564E–4

0.80

–127

100–1500a

68

82

0.3784E-5

0.71

512

0.4289E–6

0.95

–120

300–500

68

33

0.4511E–4

0.69

320

0.6612E–5

0.79

–127

35

0.1600E–5

0.65

374

0.3718E-4

0.89

–173

300–1500 67 ac = 108 µF, R ad ad

= 7 k· cm2

solution resistance (Rs ) [3]. These models describe the formation of two loops in a Nyquist diagram. A high frequency recording loop, the diameter of which depends on the potential, corresponds to the activity of the oxide film (it is determined by the serial resistance, R por e and the space charge capacitance C P E por e ). On the other hand, the low-frequency loop is related to the surface boundary of the oxide layer— solution. An Rct and C P E dl sub-circuit has been employed to describe the lowfrequency region between 11 and 0.001 Hz. Model b: In this model, the passive film is considered to have a porous structure and to show nonideal capacitive behavior. R por e is the electrolyte resistance inside the pores and C P E por e is associated with the oxide film. Rad is the charge transfer resistance of electrochemical processes taking place inside the pore and Cad is associated with an adsorption layer. Rct and C P E dl represent the resistive and nonideal capacitive behavior of the passive film (double layer) [3].

3.3 Testing the Adhesion of Microorganisms to the Surface The results of the tests are presented in Fig. 5. On the basis of the obtained tests, it was found that the number of colonies of bacteria adhered to the examined plots varies depending on the method of surface preparation. It was found that the applied Z n O layers had a beneficial effect on the number of colonies of bacteria adhered to the surface compared to the initial state. The Z n O layer on the electrochemically polished substrate was characterized by better bacteriostatic properties, where the amount of adhered bacteria was in the range of 7.5e5–1.1e6 CFU/ml.

Adhesion of Staphylococcus Aureus on Various Biomaterial Surfaces

145

Fig. 2 Examples of impedance spectra for surface layer of T i 6 Al7 N b after sandblasted: a Nyquist plot, and b Bode diagram

Fig. 3 Examples of impedance spectra for surface layer of T i 6 Al7 N b after sandblasted with Z n O layer 1500 cycles: a Nyquist plot, and b Bode diagram

Fig. 4 Electrical equivalent circuits, a passive film as a capacitor, b porous structure of passive film

146

M. Basiaga et al.

Fig. 5 Number of bacterial colonies adhered to the surface of the tested samples

4 Conclusions As part of the work, it was proposed to apply a bacteriostatic Z n O layer by the ALD method on the T i 6 Al7 N b alloy substrate. On the basis of the obtained results, its beneficial effect on electrochemical properties as well as on limiting the adherence of bacteria to the substrate was found. The conducted electrochemical tests showed an improvement in corrosion resistance compared to the samples without the layer. Moreover, it was observed that with the increase of the application temperature and the number of cycles, the resistance to pitting corrosion would increase. In turn, the impedance tests confirmed the very good barrier properties of the Z n O coating against the effect of the Ringer solution. It was found that the Z n O layer deposited at a temperature of 300 ◦ C and 1500 cycles had the most favorable electrochemical properties, which was confirmed both in potentiodynamic and impedance tests. The study of the adhesion of microorganisms to the tested surfaces also showed the beneficial effect of the applied Z n O layer, regardless of the substrate preparation. A smaller number of colonies of bacteria adhered to the surface with Z n O layers was found compared to the samples without the layer. The obtained results may constitute the basis for the development of more detailed criteria for the evaluation of the final quality of medical devices made of T i 6 Al7 N b alloy used in the skeletal system, which will ensure the required biocompatibility of the implants and contribute to minimizing the risk of postoperative complications. Acknowledgements The project was funded by the National Science Centre, Poland, allocated on the basis of the decision No. 2018/29/B/ST8/02314.

References 1. Belcarz A, et al (2009) Adhesion of Staphylococcus epidermidis cells on T i 6 Al4 V titanium alloy surfaces modified by bioceramic layers. Eng Biomat 12

Adhesion of Staphylococcus Aureus on Various Biomaterial Surfaces

147

2. Maciejewska M, Bauer M, Dawgul M (2016) Nowoczesne metody zwalczania biofilmu bakteryjnego. Poste˛py Mikrobiol 55:3–11 3. Mohammadi F, Nickchi T, Attar MM, Alfantazi A (2011) EIS study of potentiostatically formed passive film on 304 stainless steel. Electrochimica Acta 56:8727–8733 4. Papathanasiou E, Finkelman M, Hanley J, Parashis AO (2016) Prevalence, etiology and treatment of peri-implant mucositis and peri-implantitis: a survey of periodontists in the united states. J Periodontol 87:493–501 5. Pérez-Jorge C, Conde A, Arenas M, Pérez-Tanoira R, Matykina E, de Damborenea J, GómezBarrena E, Esteban J (2012) In vitro assessment of staphylococcus epidermidis and staphylococcus aureus adhesion on TiO2 nanotubes on Ti-6Al-4V alloy. J Biomed Mater Res Part A 100A(7):1696–1705 6. Phillips P et al (2010) Biofim made easy. Wounds Int 1:1–5 7. Trampuz A et al (2007) Sonication of removed hip and knee prostheses for diagnosis of infection. N Engl J Med 357:654–663 8. Trampuz A, Piper KE, Jacobson MJ, Hanssen AD, Unni KK, Osmon DR, Mandrekar JN, Cockerill FR, Steckelberg JM, Greenleaf JF, Patel R (2007) Sonication of removed hip and knee prostheses for diagnosis of infection. N Engl J Med 357:654–663 9. Tsai MT, Chang YY, Huang HL, Hsu JT, Chen YC, Wu AYJ (2013) Characterization and antibacterial performance of bioactive Ti-Zn-O coatings deposited on titanium implants. Thin Solid Films 528:143–150 10. Zhang Z et al (2019) Anti-bacterium influenced corrosion effect of antibacterial Ti-3Cu alloy in Staphylococcus aureus suspension for biomedical application. Mat Sci Eng C 94:376–384

Developing the Technology for the Production of Personalized Polylactide Plates for Bone Assemblies Reinforced with Glass Fiber ˙ Agnieszka Dubiel, Witold Walke, and Jarosław Zmudzki

Abstract Bone stabilization with bone plates is a very common orthopedic procedure. Due to the occurrence of complex fractures, fractures resulting from osteoporosis and allergic reactions, solutions in the form of traditional bone plates are not sufficient. Therefore, the use of composite materials is increasingly proposed for such fractures. The aim of this study was to obtain the appropriate mechanical properties of the proposed glass fiber-reinforced PLA-based composite using 3D technology. In order to verify the mechanical properties, the samples were subjected to a strength analysis using the tensile test and the three-point bend test. The research has shown the possibility of using reinforcements in 3D printing, the use of means to accelerate the rehabilitation of the patient after bone fusion surgery. Keywords Composite · Bone plates · Bone fractures

1 Introduction Every few seconds, bone fractures occur all over the world. Bones fracture occurs as a result of an injury, impact, or other unfortunate situation. A sedentary lifestyle and the occurrence of various diseases affecting bone structure have a negative impact on the percentage of broken bones. In addition, people with osteoporosis are much more likely to experience complicated bone fractures, e.g. with bone fragments [11]. A fracture is a serious damage that led to a break in the bone structure throughout its cross section. The research focused on a group of fractures in the distal humerus. Fractures of the humerus occur in three places. The greatest number of cases, about 40% of fractures, occurs in the proximal part, i.e. in the area of the shoulder joint. The most complicated fracture is the fracture of the humerus in the elbow joint. This case occurs as a result of an uncontrolled fall on a limb that is straightened or bent at A. Dubiel (B) · W. Walke Faculty of Biomedical Engineering, Silesian University of Technology, Zabrze, Poland e-mail: [email protected] ˙ J. Zmudzki Faculty of Mechanical Engineering, Silesian University of Technology, Gliwice, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_16

149

150

A. Dubiel et al.

Fig. 1 Presentation of the statistical distribution of fractures within the upper limbs in 2012–2018 [18]

the elbow. To show the scale of fractures in the upper limb area, statistics from the National Health Fund were carried out. According to the NHF statistics maps Fig. 1, we can analyze the number of fractures that occurred in individual provinces in the period 2012–2018. The statistics are expressed in the number of cases per 100,000 residents. In 2012 was 140.8 cases per 100 thousand inhabitants, compared to 2018 there was an increase in the reported cases to 157.6 cases per 100,000 inhabitants. In 2018 increase in fracture cases which could have been caused by the extension of life expectancy and an increase in the number of inhabitants in Poland [11, 18]. Medicine has made great technological progress, providing numerous options for the treatment, prevention and stabilization of fractures. Due to the smaller number of cases and the more complicated nature of the fractures, there are fewer solutions in the upper limb. Technological solutions in this area are limited to various metal joining plates. After applying such a solution, patients struggle with the problem of stiffening of the skeletal system after the phase of bone union [2, 6, 11, 14, 16]. The result of such a problem is the late removal of the plaque, which may result in a secondary bone fracture [7, 9, 15]. An inherent problem with the use of metal bone plates is the need for reoperation, which may expose the patient to complications during surgery and extend the rehabilitation time [9, 12]. Another problem for patients may be the occurrence of an allergic reaction or corrosion of the steel components in the organs. This applies in particular to such elements as:

Developing the Technology for the Production ...

151

Fe, Cr, Ni, Al or V. An alternative to this type of problem is the use of protective layers or coatings with increased biocompatibility or polymeric materials. The use of polymeric materials for implants in bone surgery is not always possible due to their limited load capacity. Hence, more and more often metal biomaterials are replaced with composite materials based on polymeric materials with enhanced mechanical properties [2, 5, 8, 12–14, 17]. Polymer materials show less interference with computed tomography (CT) and magnetic resonance imaging (MRI), which contributes to monitoring fracture healing using medical imaging methods. Figure 2 shows the mechanical properties of polymers and their composites that are widely used and researched as orthopedic implant materials. Composite materials can prevent reoperation, which is a significant benefit for the patient. The advantage of composites with a polymer matrix is the ability to regulate their mechanical characteristics with many variable factors depending on different applications. These variables include matrix types, types of reinforcement phases, particle size distribution of the reinforcement phase, and the number of reinforcement phases. An example of such a material is the multi-molecular PLA (polylactide), which is used to supplement bone defects, among others in the area of the skull. This material is often used in rapid prototyping technology, which enables the creation of personalized implants [1, 2, 4, 6, 7, 10, 16]. An example of such composites can be the research of Zahr et al. designed a kind of board made of carbon

Fig. 2 Mechanical properties of different materials [10]

152

A. Dubiel et al.

fiber/flax/epoxy resin composite materials. The bending strength and tensile strength of this composite are 510.6 MPa and 399.8 MPa, respectively, which are close to metallic materials, and their flexural modulus and elastic modulus are 41.7 GPa and 57.4 GPa, respectively, which are closer to cortical bone compared to traditional titanium alloys. Additive technology, or 3D printing, is more and more often used in medicine. When using biodegradable PLA, a good production method is 3D printing in FDM technology, which allows for additional personalization of the plate to the shape of the bone and the type of fracture [1, 4, 7, 10].

2 Materials 2.1 Samples Two types of samples were subjected to strength tests. The first group are plates printed with full PLA filling with a density of 0.8 g/cm3 . The second group includes plates with a PLA skeleton reinforced with glass fiber. During the three-point bending test, the composite samples were additionally tested for resin hardening. Samples for the static tensile test were designed in accordance with the PN-EN ISO 527-1:2012 standard. The samples for the three-point bending test were designed in accordance with the PN-EN ISO 178:2011 standard. The printed PLA skeleton was filled with a composite that consisted of alternating fabric layers and 0.4 mm thick PLA-printed spacers. The copositive was produced by hand lamination. The layers were joined with EPIDIAN 53 epoxy resin. The fabric layers were arranged according to the printing direction. The composite with the 25 g/mm2 fabric consisted of 5 layers of fabric and 5 PLA spacers. In contrast, the composite with fabric 110 g/mm2 consisted of 4 layers of fabric and 4 layers of PLA spacer.

2.2 Manufacturing Technology The sample design was saved in the form (*.STL format) imported into Ultimaker CURA 3D. The program decomposes the model into individual layers produced in the incremental technology. In order to increase the stability of the element, additional supports were provided in the design. The samples were printed on the Creality Ender-3 printer. Printing temperature 210 ◦ C, the bed was heated to 60 ◦ C. Nozzle 0.4 mm, filament speed 40 mm/s, retraction 3 mm. All samples were printed without the need to attach supports.

Developing the Technology for the Production ...

153

3 Methods 3.1 Tensile Test The tensile bend test was performed with the use of the ZWICK ZSN 100 testing machine. The dimensions of the samples 150 × 20 × 4 [mm]. The samples were made in accordance with the PN-EN ISO 527-1:2012 standard. The following parameters were determined in the test: the value of the breaking force, tensile stresses and elongation.

3.2 Bending Test (Three-Point) The three-point bending test was performed with the use of the ZWICK ZSN 100 testing machine. The dimensions of the samples were 64 × 10 × 3 [mm]. The spacing of the supports for the bending test was 50 mm. The samples were made in accordance with the PN-EN ISO178:2011 standard. The following parameters were determined in the test: bending force, deflection path, bending moment and bending strength.

4 Results 4.1 Tensile Test By analyzing the results for PLA samples Table 1, we can determine the correct course of the sample breaking path. There are no particular characteristic points in the graphs. The mean stress was 48.57 MPa Table 2. The composite samples showed a positive effect of the reinforcement.

Table 1 Tensile test of PLA samples Sample Fmax [N] 1 2 3 4 5

1962 1914 1910 1977 1921

σ [MPa]

. [%]

σaverage [MPa]

49.06 47.85 47.77 49.43 48.03

4.15 3.01 3.89 3.69 2.96

48.57

154

A. Dubiel et al.

Table 2 Tensile test samples as PLA skeleton + resin composite with glass fabric 25 g/mm2 Sample Fmax [N] σ [MPa] . [%] σaverage [MPa] 1 2 3 4 5

1786 1811 1791 1813 1792

52.09 51.32 52.79 51.83 48.98

4.53 4.22 4.78 4.38 4.42

Table 3 Bending test (three-point) of samples made of PLA Sample Fmax [N] f i [mm] Mg [Nmm] 1 2 3 4 5

117 119 123 125 118

12 9 8 9 10

1465.35 1496.91 1540.13 1574.57 1485.05

50.8

σf m

σaverage [MPa]

32.56 33.26 34.22 34.91 33.12

33.58

4.2 Bending Test (Three-Point) The correct course of the fracture path was demonstrated for samples with full PLA filling. There are no particular characteristic points in the graphs. The mean stress was 33.58 Mpa. The results are shown in the Table 3. When analyzing the results for the composite with glass fabric 25 g/mm2 , a characteristic small peak was noticed, indicating matrix fracture. Despite the warp breakage, the fabric fibers continued to operate, which may suggest insufficient bonding of the carcass and resin filling. Few air chambers were noted after cross-section analysis. The average stress was 45.28 MPa. The results are presented in Table 4. The analysis of the graphs gave similar results as in the previous sample, except that we can observe a longer load transfer path after the matrix fracture to the point of complete fracture. The average stress was 45.6 MPa Table 5. The situation was different for the samples where each layer was cured sequentially. The results showed good behavior of the fabric, which, after exceeding the maximum force, still played the role of reinforcement, without completely breaking the sample. Despite similar reinforcement behavior, the 110g fabric showed higher mean stresses than the lower basis weight fabric. The Tables 6 and 7 show the results of samples where each layer was cured successively.

Developing the Technology for the Production ...

155

Table 4 Bending test (three-point) of samples PLA skeleton + resin composite with glass fabric 25 g/mm2 Sample Fmax [N] f i [mm] Mg [Nmm] σf m σaverage [MPa] 1 2 3 4 5

93 89 107 104 124

9 11 8 8 9

1173.45 1116.59 1339.96 1311.93 1557.21

45.06 48.46 55.97 43.22 33.58

45.28

Table 5 Bending test (three-point) PLA skeleton + resin composite with glass fabric 110 g/mm2 Sample Fmax [N] f i [mm] Mg [Nmm] σf m σaverage [MPa] 1 2 3 4 5

108 110 109 115 128

11 11 11 10 9

1352.26 1386.44 1398.53 1462.78 1607.95

39.81 53.24 42.31 43.28 44.05

45.6

Table 6 Bending test (three-point) of samples PLA skeleton + resin composite with glass fabric 110 g/mm2 each layer hardened in turn Sample Fmax [N] f i [mm] Mg [Nmm] σf m σaverage [MPa] 1 2 3 4 5

103 79 74 92 91

17 15 11 9 13

1291.31 989.33 932.38 1157.68 1143.95

38.02 37.99 45.99 48.35 49.65

44.01

Table 7 Bending test (three-point) of samples PLA skeleton + resin composite with glass tannin 25 g/mm2 each layer hardened in turn Sample Fmax [N] f i [mm] Mg [Nmm] σf m σaverage [MPa] 1 2 3 4 5

89 89 86 98 89

9 9 8 10 10

1116.83 1116.51 1079.25 1232.31 1123.39

31.79 38.11 29.52 35.13 38.28

34.55

156

A. Dubiel et al.

5 Discussion The currently proposed materials for bone plates are based on metal alloys, which adversely affect the patient’s body during the corrosion process. In addition, the currently used plates in the healing process do not allow axial movements in the fracture fissure, which may favorably affect callus formation. Bone stiffness is a very common problem, a temporary condition after the bone plate is left in the body. The available solutions are insufficient, related to complicated fracture problems. The proposed research was based on the creation of a new material intended for the production of personalized polylactide plates for bone anastomosis reinforced with glass fiber. The material was subjected to strength tests to evaluate its strength in comparison to standard tiles. The designation of the material for bone plates for fractures in the upper limb does not transfer such loads as the material intended for e.g. hip joint prosthesis. The analysis of the results showed that the strengthening of PLA with the glass fabric increased the load capacity of the samples. The best results were obtained for the PLA + glass fabric 110 g/mm2 skeleton. Differences in the laminate production technology can be observed for the PLA composite + fabric 25 g/mm2 . In this case, the technology of laying hardened layers immediately gave a better result. The use of 3D printing can be used to create personalized bone plates appropriately suited to specific fractures. Bone plates made of polymer are not as stiff as currently used, which makes it possible to accelerate the rehabilitation of the patient after bone fusion surgery. The results showed that this is a good direction for further modification of the tested material in order to improve the mechanical properties. Additionally, in further research on this type of material, it is possible to propose a different reinforcement and a polymer matrix to achieve the expected strength values.

6 Conclusions The research was based on the development of a technology for the production of personalized polylactide plates for bone fixation reinforced with glass fiber. The presented results showed that the use of reinforcement of the polylactide skeleton with a fabric of 110 g/mm2 increased the destructive energy of the samples as a result of increased deformation. The use of fabric to strengthen the skeleton improved the load-bearing capacity of the samples produced. The analysis of the research showed the possibility of using the additive manufacturing technology to produce personalized bone implants from polylactide and the strengthening of these skeletons with glass fiber in the resin matrix.

Developing the Technology for the Production ...

157

References 1. Bagheri ZS, El Sawi I, Bougherara H, Zdero R (2015) Osteogenesis and cytotoxicity of a new Carbon Fiber/Flax/Epoxy composite material for bone fracture plate applications. Mater Sci Eng C Mater Biol Appl 46:435–442 2. Brancato V, Oliveira JM, Correlo VM, Reis RL, Kundu SC (2020) Could 3D models of cancer enhance drug screening? Biomaterials 232:119744 3. Cakir O, Subasi M, Erdem K, Eren N (2005) Treatment of vascular injuries associated with limb fractures. Ann R Coll Surg Engl 87:348–352 4. Carvalho MR, Reis RL, Oliveira JM (2018) Mimicking the 3D biology of osteochondral tissue with microfluidic-based solutions: breakthroughs towards boosting drug testing and discovery. Drug Discov Today 23:711–718 5. Choi JH, Jeon H, Song JE, Oliveira JM, Reis RL, Khang G (2018) Biofunctionalized lysophosphatidic acid/silk fibroin film for cornea endothelial cell regeneration. Nanomaterials 8(5):290 6. Ganal-Antonio AK, Samartzis D, Bow C, Cheung KMC, Luk KDK, Wong YW (2006) Disappearing bone disease of the humerus and the cervico-thoracic spine: a case report with 42-year follow-up. Spine J 2:67–75 7. Kroese-Deutman HC, Vehof JWM, Spauwen PHM, Stoelinga PJW, Jansen JA (2008) Orthotopic bone formation in titanium fiber mesh loaded with platelet-rich plasma and placed in segmental defects. Int J Oral Maxillofac Surg 37:542–549 8. Liesmäki O, Plyusnin A, Kulkova J, Lassila LVJ, Vallittu PK, Moritz N (2019) Biostable glass fibre-reinforced dimethacrylate-based composites as potential candidates for fracture fixation plates in toy-breed dogs: Mechanical testing and finite element analysis. J Mech Behav Biomed Mater 96:172–185 9. Li J, Qin L, Yang K, Ma Z, Wang Y, Cheng L, Zhao D (2020) Materials evolution of bone plates for internal fixation of bone fractures: A review. J Mater Sci Technol 36:190–208 10. Luthringer BJC, Feyerabend F, Römer R (2014) Magnesium-based implants: a mini review. Magnes Res 27(4):142–54 ´ 11. Marciniak J (2013) Biomateriały. Wyd. Pol. Sl 12. Okada M, Takamatsu K, Oebisu N, Nakamura H (2011) Reversed lateral upper arm flap with a vascularised fragment of the humerus for reconstruction of ulna shaft fracture after resection of malignant tumour: A case report. J Plast Reconstr Aesthetic Surg 64:1373–1376 13. Olsen D, Egger E, Gottsauner-Wolf F, Giurea A (1995) Biomedical application of selected polymeric materials. J Invest Surg 10:379–386 14. Ono T, Takayanagi H (2017) Osteoimmunology in bone fracture healing. Curr Osteoporos Rep 15:367–375 15. Pina S, Rebelo R, Correlo VM, Oliveira JM, Reis RL (2018) Bioceramics for osteochondral tissue engineering and regeneration. In: Advances in experimental medicine and biology. Springer, New York, pp 53–75(2018) 16. Schaller B, Saulacic N, Beck S, Imwinkelried T, Liu EWY, Nakahara K, Hofstetter W, Iizuka T (2017) Osteosynthesis of partial rib osteotomy in a miniature pig model using human standardsized magnesium plate/screw systems: effect of cyclic deformation on implant integrity and bone healing. J Cranio-Maxillofacial Surg 45:862–871 17. Siemionow M (1990) Histopathology of microarterial anastomoses: end-to-end versus end-inend (sleeve) technique. J Hand Surg Am 15:619–625 18. Map of NHF statistics (2019). https://sga.waw.pl/

Determination of the Breaking Force of Surgical Threads with the Use of a Testing Machine Anita Kajzer, Kamila Kozioł, Wojciech Kajzer, and Halina Malinowska

Abstract The aim of the study was to compare the value of the suture breaking force in the baseline state and after taking into account the exposure time to Ringer’s solution (7 and 14 d). Taking into consideration also the sewing process, the value of the force causing the thread to break away from the needle was determined. Absorbable threads (PGLA LACTIC USP 2/0 and 3/0; PGA USP 2/0 and 3/0) and non-absorbable threads (PA USP 2/0 and 3/0) were selected for the study and divided into groups according to their diameter and time of exposure to Ringer’s solution. In order to determine the value of the forces, a static tensile test was performed on the MTS Criterion Model 45 testing machine with a tensile speed of 30 cm/min. Additionally, in order to assess the values of the forces causing the breaking of the sutures made with the threads selected for testing, two samples were prepared, which differed in the number of straight knotted stitches with which the animal skin was sewn. In the first case, 9 sutures were used, in the second—11. On the basis of the obtained results, a higher value of the breaking force in the threads size USP 2/0 was found, both absorbable and non-absorbable. A lower value of the breaking force in relation to the initial state was also observed for the threads exposed to Ringer’s solution. In the case of sewn animal skin, it can be concluded that the highest value of the force was obtained for tearing 11 straight knotted sutures made with USP 2/0 nylon threads (F = 26.1 N). On the other hand, the smallest force (F = 17 N) was obtained to tear 9 straight knotted sutures made with the same threads, size USP 3/0. Keywords Absorbable and non-absorbable surgical suture · Mechanical properties

A. Kajzer (B) · K. Kozioł · W. Kajzer Department of Biomaterials and Medical Devices Engineering, Faculty of Biomedical Engineering, Silesian University of Technology, ul. Roosevelta 40, 41-800 Zabrze, Poland e-mail: [email protected] H. Malinowska YAVO Sp. z. o. o, ul. Bawełniana 17, 97-400 Bełchatów, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_17

159

160

A. Kajzer et al.

1 Introduction Surgical threads (Fig. 1) have been used in medicine for several hundred years for tissue fixation [1]. Both for those which are injured as a result of trauma and after surgical operations [2, 3]. They have been classified as medical devices and defined in Directive 93/42/EEC [4]. Their application for suturing a wound (Fig. 2) is about bringing the damaged tissue closer so that it can regenerate itself. The basic mistake is to understand this process as the mechanical maintenance of tissue continuity [5]. Depending on the wounds, appropriately selected surgical threads are used, allowing faster and more effective regeneration of the damaged site. The type of material used, its surface, size, diameter, structure and biodegradation are of great importance when performing surgical procedures. The basic division is whether a given surgical thread is classified as absorbable material in contact with the human body or not. This division is one of the most appropriate for the classification and selection of a suture during a surgical procedure [1].

Fig. 1 Example of a surgical thread [11]

Fig. 2 Examples of types of knots for sewing wounds [12]

Determination of the Breaking Force of Surgical Threads ... Table 1 Selected YAVO surgical sutures Type of thread Representative Size [USP] Absorbable Absorbable Absorbable Absorbable Non - absorbable Non - absorbable

PGLA LACTIC PGLA LACTIC PGA PGA PA PA

2/0 3/0 2/0 3/0 2/0 3/0

161

Diameter [mm]

Length [cm]

0.3 0.2 0.3 0.2 0.3 0.2

90 90 90 90 90 90

2 Materials and Methods Absorbable and non-absorbable threads were selected for the study. The parameters of the threads used are presented in Table 1. In the first stage, absorbable sutures without a needle were exposed to Ringer‘s solution for 7 and 14 d necessary for the regeneration of soft tissue in the human body. The samples prepared in this order were placed in a Q-CELL laboratory incubator at the temperature of 36.6 ◦ C. Non-absorbable threads were tested (Fig. 3A) in a straight knot (Fig. 3B) and with a needle, comparing the breaking force value for the sizes USP 2/0 and USP 3/0. Additionally, three sewn animal skins (Fig. 4A, B)of the same dimensions were prepared. Two samples had eleven straight stitches, one of which was sewn with USP 2/0 nylon threads, the other with the same type of thread, USP 3/0 size. Additionally, the skin was stitched with nine 3/0 size sutures. The static tensile test was carried out on the MTS Criterion Model 45 testing machine (Fig. 3A) in accordance with the recommendations of PN-EN ISO5271: 2012 [6] and the Polish Pharmacopoeia, 11th edition, with a tensile speed of 30 cm/min [7]. Each thread was fixed so that its measuring part had the same distance from the handles (150 mm), and in the case of testing the thread breaking force, particular care was taken to ensure that the straight knot was in the middle (Fig. 3B).

3 Results The obtained results of nylon, PGLA LACTIC and PGA threads, sizes USP 2/0, 3/0, taking into account the condition and exposure to Ringer’s solution are presented in Table 2. The values of the force detaching the thread from the needle for PA, PGLA LACTIC and PGA threads are presented in Table 3. Additionally, exemplary results are presented in the force-elongation charts for respectively selected threads (Figs. 5 and 6). Table 4 shows the maximum value of the force needed to tear the single straight knotted sutures with which the animal skin was sewn.

162

A. Kajzer et al.

Table 2 Test results for individual threads in the straight knot Surgical thread, size, condition, straight knot Average value of the breaking force, F [N] PGLA LACTIC USP 2/0, initial state PGLA LACTIC USP 3/0, initial state PGA USP 2/0, initial state PGA USP 3/0, initial state PGLA LACTIC, USP 2/0, 7 days - Ringer’s solution PGLA LACTIC, USP 2/0, 14 days - Ringer’s solution PGLA LACTIC, USP 3/0, 7 days - Ringer’s solution PGLA LACTIC, USP 3/0, 14 days - Ringer’s solution PGA, USP 2/0, 7 days - Ringer’s solution PGA, USP 2/0, 14 days - Ringer’s solution PGA, USP 3/0, 7 days - Ringer’s solution PGA, USP 3/0, 14 days - Ringer’s solution PA, USP 2/0, initial state PA, USP 3/0, initial state

37.33 ± 1.25 25 ± 1.00 45 ± 1.73 24.33 ± 0.76 32.33 ± 2.49 28.17 ± 3.57 21.67 ± 1.59 20.83 ± 0.86 37.67 ± 1.79 31.83 ± 2.03 22.83 ± 1.77 17.5 ± 1.50 32.33 ± 2.21 19.33 ± 1.59

Table 3 Test results for individual threads, thread-needle connection Surgical thread, size Average value of the force (F) of the tested threads [N] 26.17 ± 2.57 19.17 ± 2.11 31.17 ± 3.24 17.67 ± 2.29 28.83 ± 3.34 17.17 ± 3.53

PA, USP 2/0 PA, USP 3/0 PGLA LACTIC, USP 2/0 PGLA LACTIC USP 3/0 PGA, USP 2/0 PGA, USP 3/0

Table 4 Test results for USP 2/0, 3/0 nylon threads with which animal skin was sewn Force F [N] - 11 straight Force F [N] - 11 straight Force F [N] - 9 straight knotted knotted sutures USP 2/0 knotted sutures USP 3/0 sutures USP 3/0 26.1

18.4

17.1

Determination of the Breaking Force of Surgical Threads ...

163

Fig. 3 Stand for testing the breaking force: A MTS testing machine, B example of a thread fixed in the holders

Fig. 4 Animal skin sewn together with nylon threads: A before the test, B after the test

4 Discussion The aim of the study was to determine the effect of the exposure time to Ringer’s solution on the value of the breaking force of the threads selected for testing. In addition, force of breaking the thread from the needle was assessed for USP sizes 2/0 and 3/0 for both absorbable and non-absorbable sutures. Additionally, the animal skin was sewn with 9 and 11 straight knotted sutures using USP 2/0 and 3/0 nylon threads. On the basis of the obtained results it can be concluded that the force causing nylon threads to break in the straight knot size USP 2/0 (F = 32.33 N) is 1.67 times

164

A. Kajzer et al.

greater than for the same threads in size USP 3/0 (F = 19.33 N). Also, a higher value of the needle thread peel force was obtained for USP 2/0 (F = 26.17 N) nylon threads than for USP 3/0 (F = 19.17 N) threads. In the case of absorbable PGLA LACTIC multifilament threads of USP 2/0 size, it was found that the suture breaking force (F = 37.33 N) in the straight knot is one and a half times greater than for the same threads of USP 3/0 size (F = 25 N). It was also noticed that the force value for PGLA LACTIC USP 2/0 thread exposed to Ringer’s solution during 7 d (F = 32.33 N) decreased by 5 N. The force value obtained for the same thread exposed to Ringer’s solution during 14 d (F = 28.17 N) is 1.33 times less than the initial force values. Thus, it can be noticed that the value of the force in relation to the initial state decreases with the increase of the residence time of the thread in the Ringer’s solution. The force obtained for the PGLA LACTIC USP 3/0 thread in its initial state (F = 25 N) is by 3.33 greater than the force value for the same thread exposed to Ringer’s solution for 7 d (F = 21.67 N) and by 4.17 greater than the force value for the thread remaining in the Ringer’s solution for 14 d. In this case, one can also see the dependence of the decreasing value of the breaking force with the increase of the exposure time to the Ringer’s solution. The value of the force obtained for the PGA USP 2/0 multifilament threads (F = 45 N) is 1.85 times higher for the same threads in the USP 3/0 size (F = 24.33 N). It is worth noting that also for PGA USP 2/0 suture the force value decreases depending on the exposure time to the Ringer’s solution. This value compared to the initial state of the PGA USP 2/0 thread (F = 45 N) is 1.2 times lower (F = 37.67 N) in the case of exposure to Ringer’s solution for 7 d and 1.41 times lower (F = 31.83 N) - 14 d. The force value obtained for the PGA USP 3/0 thread in its initial state (F = 24.33 N) is 1.5 greater than the force value for the same thread exposed to Ringer’s solution (F = 22.83 N) and 6.83 N greater than the force value for the thread remaining in the Ringer’s solution for 14 d. Higher value of the

Fig. 5 Sample diagrams of PA USP 3/0, 2/0 static thread tensile test in a straight knot

Fig. 6 Examples of PGLA USP 3/0, 2/0 static thread tensile test charts in a straight knot

Determination of the Breaking Force of Surgical Threads ...

165

needle thread detachment force was obtained for the PGLA LACTIC thread for the USP 2/0 size (F = 31.17 N) than for USP 3/0 (F = 17.67 N). The same relationship can be seen in the case of PGA threads. The value of the needle thread peeling force obtained is 1.7 times greater for USP 2/0 (F = 28.83 N) than USP 3/0 (F = 17.17 N). It is noteworthy that the obtained force value for PGA USP 3/0 (F = 17.17 N) and PGLA LACTIC USP 3/0 (F = 17.67 N) threads is very close. However, in the case of PGA USP 2/0 (F = 28.83 N) and PGLA LACTIC USP 2/0 (F = 31.17 N) threads, the difference in value is 2.34 N. The value of the breaking force in the straight knot of the PGA USP 2/0 thread. The (F = 45 N) in the initial state is 1.21 times greater than for PGLA LACTIC USP 2/0 thread (F = 37.33 N). On the other hand, the value of the breaking force of the PGA USP 3/0 thread (F = 24.33 N) is very close to the value of the force obtained for the PGLA LACTIC USP 3/0 thread (F = 25 N). In the case of the results obtained with the USP 2/0 nylon threads, with which the animal skin was sewn together, it can be seen that the highest value of the force was obtained for tearing 11 straight knotted sutures (F = 26.1 N). This value is 1.41 times greater than when the same number of seams sewn with PA USP 3/0 (F = 18.4 N) are torn. The breaking force value of a seam made on animal skin sewn with 9 knotted stitches with PA USP 3/0 thread (F = 17.1 N) is 1.294 less than the force value of USP 2/0 nylon thread sewn with 11 stitches. In the presented diagrams (Fig. 6) of absorbable multifilament threads - PGLA LACTIC, a decrease in the force value is observed as a result of breaking subsequent fibers. On the other hand, in the case of nylon threads - non-absorbable, single-filament threads, an increasing force value is observed in the graphs of the static tensile test until it is completely broken (Fig. 5). The authors of the article [8] also examined PGA absorbable threads and USP 2/0 nylon sutures. These threads were stretched at a speed of 30 mm/min with a length of 15 cm for the knot test and 12 cm to determine the force of needle-thread separation. The value of the breaking force of multifilament absorbable PGA threads in the straight knot in the initial state was 35 N. The thread was also exposed to Ringer’s solution for 7 and 14 d, then the force value gradually decreased and the following results were obtained: 23 and 16 N. The PGA thread was subjected to also examining the detachment of the thread from the needle. In this case, the force value was 45 N. Additionally, the nylon threads in the straight knot and the thread separation from the needle were tested. The values of the obtained forces were respectively 29 and 26 N. The authors also carried out force tests of ordinary knotted sutures, with which animal skin with dimensions of 110 × 60 mm was sewn. The stretching speed was 4 mm/min. It is worth noting that although the authors [8] obtained a higher value of the force needed to tear a smaller number of seams (F = 41 N), a much lower stretching speed was used. In the authors’ own research this speed was 300 mm/min, it was 75 times higher than in the research by M. Zurek et al. [8]. On the other hand, the authors of the article [9] examined surgical threads in sizes USP 2/0 and USP 3/0. The tensile speed used in the test was 100 mm/min, in addition, a single knotted seam was used and the place of tearing was observed. The breaking force of surgical sutures was also investigated by the authors [9]. The values obtained were in the range of 96 to 102 N for the USP 2/0 PGA thread and from 53 to 75 N for the same USP 3/0 thread. In addition, the value of the force obtained for the nylon threads in the USP 2/0 size was from 63 to

166

A. Kajzer et al.

74 N, and for the USP 3/0 threads from 22 to 42 N. These values are much higher than those obtained in our own research. The value of the force obtained for the USP 3/0 nylon threads (F = 22 N) is close to the force value obtained in the own research for the same threads of the same size (F = 19.33 N). On the other hand, the authors [10] examined nylon threads available in the United States. Nylon threads from seven different manufacturers in the size USP 5/0 were tested. The threads were stretched at a speed of 125 mm/min. It can be seen that the value of the obtained force needed to break the thread in the article for different manufacturers is similar. Moreover, on the basis of own research and the presented article, it was noticed that the force value decreased with a smaller diameter. Based on own research and the presented articles [8–10], it can be concluded that the value of the force obtained both for the thread with a straight knot and for the breaking of the thread-needle connection depends on several factors. The first basic is the thread type or diameter. Next, the environment in which the tested thread stays and with which it is stretched by the speed. Another important aspect is the test conditions and the manufacturer.

References 1. Zapalski S, Che˛ci´nski P (1999) Szwy chirurgiczne. Alfa Medica Presss, Bielsko Biała 2. Zie˛bowicz A, Kajzer A, Kajzer W, Marciniak J (2010) Metatarsal osteotomy using doublethreaded screws - biomechanical analysis. In: Conference on information technologies in biomedicine. Advances in soft computing, vol 69. Springer, Heidelberg, pp 465-472 3. Basiaga M, Walke W, Staszuk M, Kajzer W, Kajzer A, Nowiska K (2017) Influence of ALD process parameters on the physical and chemical properties of the surface of vascular stents. Arch Civ Mech Emg 17:32–42 4. Council Directive 93/42/EEC of 14 June 1993 relating to medical devices 5. Bielecki K (2008) Narze˛dzia, protezy i szwy chirurgiczne. Lublin, Makmed, vol 2 6. Polish Committee for Standardization: PN-EN ISO527-1: 2012 (pol.) (2013) Plastics - determination of mechanical properties under static stretching - part 1: general principles, Catalog of Polish Standards 7. Polish Pharmacopoeia XI edition, 1 (2017) 11431-1440 8. Zurek M, Kajzer A, Basiaga M, Jendru´s R (2016) Wła´sciwo´sci wytrzymało´sciowe wybranych polimerowych nici chirurgicznych. Polimery 61(5):334–338 9. Wesołowski P., Bakuniak P, Iwanowski K, Kresa I, Wojtowicz A (2012) Porównanie wytrzymało´sci mechanicznej wybranych nici stosowanych w chirurgii stomatologicznej. Nowa Stomatologia 10. Callahan TL, Lear W, Kruzic JJ, Maughan CB (2016) Mechanical properties of commercially available nylon sutures in the United States. Wiley Periodicals 11. Nici chirurgiczne ATRAMAT Nylonowe. www.skle.meringer.pl, Accessed 01 Dec 2019 12. Nawceniak-Balczerska M Szycie pojedynczej rany skóry i tkanki podskórnej, www. zaufanekliniki.pl, Accessed 17 Nov 2019

Study of Physical Properties of Additively Manufactured and Post-processed 3D Porous Structures Intended for Implants Wojciech Kajzer, Katarzyna Gieracka, Mateusz Pawlik, Marcin Kaczmarek, and Anita Kajzer

Abstract The work aimed to determine the effect of finishing porous threedimensional structures manufactured with additive technologies such as Selective Laser Sintering (SLS) and Fused Deposition Modeling (FDM) on their physical properties. The work focuses on the influence of the material, type of surface treatment as well as the thickness and height of the strut forming the tested structures on their surface quality, the degree of porosity, and compressive strength. Moreover, the compatibility of the 3D printed porous structure to the developed 3D model designed in the CAD software was also analyzed. The test samples were made of polyamide PA12 (SLS technology) and PLA (FDM technology). Depending on the 3D printing technology used, the impact of different types of finishing was assessed. The scope of the research included: mass measurement, determination of geometrical features, porosity examination by gas pycnometry, macroscopic observations, and static compression test. The analysis of the obtained results showed that with the use of SLS technology, the 3D physical model is characterized by greater accuracy compared to the model manufactured in FDM technology. It was also found that the smaller the strut height and thickness of the porous 3D structure, the more compressive strength the structure has. In the case of finishing 3D printing, it was found that for SLS technology, sandblasting with glass beads is the optimal treatment, while for models printed by the FDM method - annealing. Keywords 3D porous structures · Post-processing operation for 3d printing · Additive technology · SLS - Selective Laser Melting · FDM - Fused Deposition Modeling · Mechanical research · Macroscopic W. Kajzer (B) · M. Kaczmarek · A. Kajzer Department of Biomaterials and Medical Devices Engineering, Faculty of Biomedical Engineering, Silesian University of Technology, ul. Roosevelta 40, 41-800 Zabrze, Poland e-mail: [email protected] K. Gieracka Science Club SYNERGIA, Department of Biomaterials and Medical Devices Engineering, Faculty of Biomedical Engineering, Silesian University of Technology, Zabrze, Poland M. Pawlik CABIOMEDE Sp. z o.o., Kielce, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_18

167

168

W. Kajzer et al.

1 Introduction Tissue scaffolds enable the regeneration of damaged tissue by restoring its normal vital functions. In tissue engineering, they play the role of a structural matrix, corresponding to, among others, differentiation, multiplication, and attachment of cells. They also ensure proper maintenance of tissue cells in the newly formed structure [1–3]. The wide possibilities of tissue scaffolds have influenced their various applications, including medicine and dentistry, where they are currently used to reconstruct damaged structures [1]. Tissue engineering puts high demands on scaffolds to recover and restore the functionality of the reconstructed structure as accurately as possible. In the context of medicine and dentistry, the selection of material as well as the technology of producing porous structures depends on many factors, including the place of implantation, the type of reconstructed tissue, etc. On the other hand, the quality and properties of porous structures are influenced by four main factors - the shape and size of the pores, porosity, and connections between them [4, 5]. In the process of designing an implant with a scaffold structure, appropriate mechanical properties should also be taken into account, adjusted to the size of the damage. This approach will avoid, among others. stress concentrations and will meet the relevant requirements of anatomical structures [6, 8, 10]. These factors are the most important considerations when designing scaffolds. Other factors, such as sterilization method, the selection of which depends on the material, degradation time depending on the regeneration rate of the damaged structure, or adhesive properties affecting cell adhesion, should also be taken into account [9]. Currently, there are many known methods of producing scaffolds, which can be divided into two groups: traditional (e.g. gas foaming, emulsification) and incremental technologies (e.g. Fused Deposition Modeling, Selective Laser Sintering). Because 3D printing gives much more possibilities, resulting from the direct production of 3D porous models designed in CAD software, nowadays this technology is dominant in the manufacturing of scaffolds [2]. The 3D printing process begins with the design of a 3D model in a CAD program. At this stage, the parameters are selected, i.e. the degree of porosity, shape, and size of the pores. The parameters of 3D printing are also adjusted, depending on the selected additive technology (including the shape of the filling) [3]. After printing the physical object, the final stage is finishing, the so-called post-processing. This stage is as important as the previous stages, because an improperly cleaned structure made of the base material (e.g. from unsintered powder, from filament threads) may cause abnormal tissue growth or deterioration of the encrustation process. Depending on the material used to produce 3D structures, various methods and means are used to clean them, such as an ultrasonic cleaner, compressed air, or solvents [7]. Currently, efforts are being made to avoid the potential risks associated with the process of printing porous models and to choose the optimal one, ensuring the best cleaning of the structure from unprocessed batch material, among the available post-processing methods. Therefore, the work aimed to choose the best method of the post-processing for porous spatial structures produced using SLS and FDM technology. In particular, the study determined the impact of the proposed finishing treatments on the

Study of Physical Properties of Additively Manufactured and Post-processed ...

169

surface morphology, porosity, mass, and compressive strength of the manufactured 3D objects. Moreover, the quality of the model designed in the CAD program was analyzed concerning the actual printed 3D object, depending on the applied additive manufacturing technology.

2 Materials and Methods Porous samples, printed with additive technologies: SLS (made of PA12 - DuraForm ProX PA) and FDM (made of PLA - Poly (L-lactide) RESOMER Filament L D1.75) were used in the research. When designing porous 3D models, two parameters were taken into account: thickness and height of the beam forming the structure - Fig. 1. Additionally, depending on the type of base material, the three most frequently used types of treatments after the 3D printing process were analyzed. After the finished stages of the 3D printing process, each sample was sterilized by steam sterilization using a cycle designed for plastics at a temperature of 121 ◦ C. The research material is summarized in Table 1.

Table 1 Research material Additive technology SLS Material Beam height [mm] Beam thickness [mm] Post-processing

PA12 0,9 1,0 0,6 0,7 0,8 0,6 0,7 IA - Initial state IB - Sandblasting IC - Acid ID - Ultrasonic cleaner

FDM

0,8

Polylactide (PLLA) 0,9 1,0 0,6 0,7 0,8 0,6 0,7 IIA - Initial state IIB - Rinsing IIC - Heating IID - Ultrasonic cleaner

Fig. 1 Model of the porous structure 0.9 mm × 0.6 mm: a geometry, b section

0,8

170

W. Kajzer et al.

The samples were divided into two main groups (depending on the manufacturing technology), and then into four subgroups (depending on the applied post-processing method). For porous structures manufactured by SLS technology, they are as follows: – samples after the 3D printing process, standard powder removal from the pores: sandblasting, pushing the pores with a needle, shaking out - initial state; – samples after the 3D printing process, sandblasting with glass microbeads of the fraction 150–250 µm (5 min) with the pressure of 4 bar; – samples after the 3D printing process, rinsing in formic acid and dichloromethane for 60 s (50:50 concentration of pure DMMA and 85% acid), then rinsing in H2 O (15 min) and drying in the open air; – samples after the 3D printing process, ultrasonic cleaning in H2 O for 30 min, then rinsing in a disinfectant (10 min), rinsing in H2 O (5 min), and drying in the open air. For a group of samples printed with the FDM method: – samples after 3D printing - initial state; – samples after the 3D printing process, rinsing in a disinfectant (10 min), rinsing in H2 O (5 min), and drying in the open air; – samples after the 3D printing process, heating at 80 ◦ C for 16 h; – samples after the 3D printing process, ultrasonic cleaning in H2 O (30 min), drying in the open air.

2.1 Determination of Geometry The measurement of the printed model geometry (i.e. length, height, and width) was performed using a caliper with an accuracy of 0.01 mm. The purpose of the measurement was to compare the actual dimensions of 3D structures to the dimensions of 3D models designed in the CAD program. For each variant, the apparent volume (V p ) necessary for the calculation of the degree of porosity was calculated. Additionally, the obtained values were compared to the volume of the 3D model designed in the CAD program.

2.2 Macroscopic Observations Macroscopic observations were carried out using the Zeiss SteREO Discovery V8 stereoscopic microscope to determine whether there was any remaining base material in the pores that was not used in the printing process (i.e. unsintered powder or filament threads). The test was performed at magnifications: 4.8×, 6×, 12×, in two perspectives, i.e. the upper and the side - Fig. 2. Attention was also paid to the occurrence of possible design errors (e.g. edge distortions, material losses, etc.) in the visible walls of construction.

Study of Physical Properties of Additively Manufactured and Post-processed ...

171

Fig. 2 Two observation perspectives of porous 3D structures

2.3 Mass Measurement To determine the impact of finishing the 3D printed object on its mass, a randomly selected sample of a given variant was weighed using an EXPLORER balance with an accuracy of 0.001 g.

2.4 Structure Porosity Testing Using Gas Pycnometry An automatic gas pycnometer ULTRAPYC 1200 e was used to determine the degree of porosity. The test time for a single sample was 10 min. There was appointed, among others actual volume (Vr z ) of the analyzed variant. To determine the percentage value of the degree of porosity, Formula No. 1 was used. The apparent volume (V p ) was used from the previously performed study (Sect. 2.1). P = (1 −

Vr z ) × 100% Vp

where: – P—porosity [%]; – Vr z —actual volume [cm3 ]; – V p —apparent volume [cm3 ]

2.5 Compression Test The static compression test was carried out with the use of the MTS Criteron 45 testing machine. The test was performed with the compression rate mm/s. The test was carried out until the test sample was destroyed. The machine was connected to a computer with the MTS TestRun Suit software. The maximum destructive force

172

W. Kajzer et al.

Pcmax was determined. If the graph did not show the characteristic curve elevation, the force PcL was also determined when the traverse of the testing machine was displaced by L = 5 mm.

3 Results and Discussion 3.1 Results of Geometry Measures The results of the average apparent volume of the 3D printed structures together with the corresponding value of the CAD model volume are presented in Fig. 3. The analysis of the results showed that additive manufacturing technologies are burdened with some dimensional inaccuracy, especially the FDM method. It was also noted that samples produced with the powder technology - i.e. SLS - are characterized by the apparent volume value closer to the modeled value.

3.2 Macroscopic Observations Exemplary results of macroscopic 3D structures made of a beam 0.9 mm high and 0.6 mm thick are shown in Figs. 4 and 5. When performing a visual assessment, in terms of the lowest content of deposited powder in the pores, it was found that the optimal type of post-processing of porous structures printed from PA12 powder is sandblasting with micro glass beads. After treatment by acid rinsing, the amount of material deposited in the pores was adversely increased, while the third method (i.e. ultrasonic cleaner) did not sufficiently clean the pores of residual building material.

Fig. 3 Comparison of the apparent volume of CAD models with samples produced by SLS and FDM methods for models with a beam height: a 0.9 mm, b 1.0 mm

Study of Physical Properties of Additively Manufactured and Post-processed ...

173

Fig. 4 The surface of PA12 samples with a beam height of 0.9 mm and a thickness of 0.6 mm, area 6×: IA Baseline IB after sandblasting IC after acid rinse ID after ultrasonic cleaning

Fig. 5 The surface of PLA samples with a beam height of 0.9 mm and a thickness of 0.6 mm, area 6×: IIA initial state, IIB after rinsing, IIC after annealing, IID after ultrasonic cleaning

When analyzing the influence of individual finishing treatments on the quality of porous structures printed with the FDM method, it was found that the best pore cleaning effect was obtained by heating and using an ultrasonic cleaner. It was also noticed that the outer surfaces of the samples subjected to the ultrasonic bath were smoother. The finishing treatment by rinsing in the disinfectant practically did not affect the elimination of filament threads from the pores.

3.3 Results of Mass Measurements The results of mass measurements of 3D porous structures printed with the SLS technology are shown in Fig. 6, while the samples manufactured by the FDM method in Fig. 7. The greatest mass reduction for structures made with powder technology was obtained after acid treatment. Since the fact that the differences in mass obtained

174

W. Kajzer et al.

Fig. 6 Average mass values of samples printed with the SLS method, with the beam height: a 0.9 mm, b 1.0 mm

Fig. 7 Mass of the samples printed by FDM method, with the beam height: a 0.9 mm, b 1.0 mm

for this method differ slightly from the results of the other two methods, a strength test should be carried out to check whether the acid treatment has not reacted with the building material, causing both mass loss and reduced strength of the entire structure. The other two methods of processing (sandblasting and ultrasonic cleaning) contributed to a similar mass loss. Taking into account the results of samples printed with FDM technology from PLA material, it was found that the greatest mass loss was achieved after annealing, whereas the smallest mass loss was observed for rinsing in the disinfectant. When analyzing the obtained results, it can be generally stated that finishing treatment (in most cases) reduces the mass of the tested porous structures, regardless of the applied manufacturing technology. This is caused by the removal of the base material that is not used during printing.

3.4 Structure Porosity Testing Using Gas Pycnometry Exemplary results of the porosity test for structures made of a beam with a height of 0.9 mm (for both printing methods) are shown in Fig. 8.

Study of Physical Properties of Additively Manufactured and Post-processed ...

175

Fig. 8 Results of the porosity degree of 3D models with a beam height of 0.9 mm, taking into account the calculated porosity of the CAD model, printed by the method: a SLS, b FDM

Fig. 9 Summary of mean values of Pcmax forces obtained from the static compression test of the samples manufactured by: a SLS; b FDM

For the majority of porous 3D structures, an upward trend in the degree of porosity was observed after the applied finishing treatment, compared to the porosity of the models in the initial state. For this reason, the variants with a downward trend in the porosity value were not taken into account in the final conclusions. It was found that acid rinsing is the best finishing treatment (increasing the porosity value) of the models produced with the SLS technology among the analyzed models, while for the models produced in the FDM technology - heating and ultrasonic cleaning.

3.5 Results of the Static Compression Test Exemplary results from the performed static compression test of porous spatial structures are shown in bar charts of the maximum compressive force with the reference to the size of the beam (Fig. 9). Selected graphs of two materials, PA12 and PLA, are presented in Fig. 10. Based on the results of the static compression test of samples made of PA12 in SLS technology, it is not possible to unequivocally state the influence of the applied

176

W. Kajzer et al.

Fig. 10 Selected graphs of the static compression test of 3D objects, two manufacturing technologies, with a beam height of 1 mm and a beam thickness of 0.8 mm

Fig. 11 Macroscopic photos after a static compression test of samples with a beam height of 0.9 mm and a thickness of 0.6 mm, area 4.8×: a PA12, top view, b PA12, side view, b PLA, top view, type I, d PLA, side view, type I, e PLA, type II

finishing treatments on their compressive strength (no repeatability in the recorded values of Pcmax forces). On the other hand, for most variants of samples made of PLA (FDM technology), a negative effect of finishing treatments was found, which reduced the compressive strength. Depending on the applied processing of the 3D model, the material could react with the chemical agent used to clean the model or the structure degraded due to excessively high temperature. Exemplary samples after compression are shown in Fig. 11. For each variant made of PA12, there was a slip between the incremental layers. Two types of cracks were observed for the structures printed from PLA: I - the fracture of the crack ran through the center of the sample (Fig. 11c and d), II - the fracture was visible at its base (Fig. 11e). The course of cracks is related to the structure of the material, i.e. in the case of samples printed with PA12 (i.e. powder), after reaching the maximum destructive force, the structure strengthened with a further displacement of the traverse of the testing machine (crack and slip ran along the grains), while in

Study of Physical Properties of Additively Manufactured and Post-processed ...

177

variants made of PLA material - the crack ran parallel between two layers, which resulted in a sudden break in the continuity of the material.

4 Conclusion Based on the conducted research, it was found that the best of the proposed treatments for finishing the porous 3D structure (influencing with the highest degree the removal of unprocessed building material from the pores, and obtaining the lowest mass and the highest porosity while maintaining good compressive strength, regardless of the geometry in relation to structures in the final state) are: – for SLA technology: sandblasting with glass microbeads; – for FDM technology: heat treatment or application of an ultrasonic cleaning. The conducted research allowed to draw the following conclusions and observations: – treatments finishing porous three-dimensional structures, produced in SLS and FDM technologies, increase the degree of porosity and mass reduction, as well as slightly lower their compressive strength compared to the initial state. None of the applied treatments fully removed the embedded building material in the pores; – the disadvantage of 3D printing methods is the production of models with dimensional inaccuracy in relation to the dimensions designed in the CAD program; it is easier to reflect the geometry of a larger structure while obtaining porosity similar to the value of the designed CAD model; – more accurate mapping of the geometric features of CAD models is obtained using the Selective Laser Sintering method; – in the case of porous structures printed with the FDM method, apart from the selection of the finishing method, attention should be paid to the appropriate selection of printing parameters, i.e. process temperature, cooling, layer thickness, filament feed rate, etc.; – the compressive strength of porous 3D structures depends on the geometry. It has generally been found to be proportional to the thickness of the beam building the structure and inversely proportional to its height. Acknowledgements The work has been financed from research project no: 07/020/B K _20/0059.

References 1. Laska A (2017) Biomateriały stosowane w in˙zynierii tkankowej do regeneracji tkanek. Zeszyty ´ Naukowe Towarzystwa Doktorantów UJ Nauki Scisłe 14(1):187–196 2. Zhang L, Yang G, Johnson BN, Jia X (2019) Three - dimensional (3D) printed scaffold and material selection for bone repair. Acta Biomater 84:16–33

178

W. Kajzer et al.

3. Dobrza´nski LA, Dobrza´nska-Dankiewicz AD, Kroll TG (2017) Metalowe materiały mikroporowate i lite do zastosowa´n medycznych i stomatologicznych. Sci Int J World Acad Mater Manuf Eng. VII:1–580 4. Mucha M, Tylman M (2013) Wielofunkcyjne biopolimerowe skafoldy jako implanty ko´sci. Eng Biomater 118:12–17 5. Yang S, Leong KF, Du Z, Chua CK (2001) The design of scaffolds for use in tissue engineering. Part I. Traditional factors. Tissue Eng 7(6):679–89 6. Zie˛bowicz A, Kajzer A, Kajzer W, Marciniak J (2010) Metatarsal osteotomy using doublethreaded screws - biomechanical analysis. In: Conference on information technologies in biomedicine. Advances in soft computing, vol 69. Springer, Heidelberg, pp 465-472 7. Kajzer W, Krauze A, Kaczmarek M, Marciniak J (2008) FEM analysis of the expandable intramedullary nail. In: Conference on information technologies in biomedicine. Advances in soft computing, 16–18 June 2008, vol 47. Springer, Heidelberg, pp 537-544 8. Kruk A, Gadomska-Gajadhur AA, Sebal A, Ru´skowski P (2017) Rusztowania tkankowe w in´zynierii tkankowej. Wyroby medyczne 4:31–35 9. Yan Y, Chen H, Zhang H, Guo C, Yang K, Chen K, Cheng R, Qian N, Sandler N, Zhang YS, Shen H, Qi J, Cui W, Deng L (2019) Vascularized 3S printed scaffolds for promoting bone regeneration. Biomaterials 190–191:97–110 10. Dobrza´nski LA Biomaterials in regenerative medicine. https://www.intechopen.com/books/ biomaterials-in-regenerative-medicine/introductory-chapter-multi-aspect-bibliographicanalysis-of-the-synergy-of-technical-biological-and-;

Impact of 3D Printing Materials on Bone Phantom Features Marta Kiel-Jamrozik, Wojciech Jamrozik, Mateusz Pawlik, and Jakub Goczyla

Abstract Bone drilling is a common orthopedic procedure used to produce drill holes for screw insertion to repair fractured bone. Hole quality depends not only on the drill geometry but also on the cutting parameters, drilling force, and drilling technique, therefore, skilled personnel are required to successfully perform bone drilling. The drilling process should limit the generation of heat (thermal energy) to avoid osteonecrosis caused by interrupted blood supply to bone tissue. Studies were performed to determine the optimal material type and infill percentage for fabricating bone phantoms via 3D printing. Phantoms made from PA, PLA, and PET were produced using different infill percentages. Our results show large differences in quality among certain materials and demonstrate that 3D printing can be successfully used for manufacturing bone phantoms. Keywords 3D printing · Bone phantom · Thermovision

1 Introduction Bone drilling is a basic technique used in surgical implantation procedures. The bone drilling process produces frictional heat, mainly between the tool and the bone, which is partially dissipated by the flow of blood and tissue fluid and also carried away by the bone chips produced. Due to the poor thermal conductivity of bone, increases in temperature can be significant. Moreover, densities of individual structural elements in the layered bone tissue structure are different, resulting in higher M. Kiel-Jamrozik (B) · J. Goczyla Faculty of Biomedical Engineering, Department of Biomaterials and Medical Devices Engineering, Silesian University of Technology, Zabrze, Poland e-mail: [email protected] W. Jamrozik Faculty of Mechanical Engineering, Department of Fundamentals of Machinery Design, Silesian University of Technology, Gliwice, Poland M. Pawlik CABIOMEDE Sp. z o. o., Kielce, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_19

179

180

M. Kiel-Jamrozik et al.

temperatures during drilling in bone with higher density, such as cancellous bone. An increase in temperature at the drilling site can cause thermal bone necrosis. Furthermore, the thermal processes leading to bone damage and necrosis can affect proper tissue healing and bone regeneration and have therefore become a major topic of scientific research. These complications can significantly slow the recovery of the patient following an operation [2, 5, 8, 9]. The critical temperature for necrosis is a controversial topic that is often discussed in the literature. Many scientists assume a critical temperature of thermal bone necrosis of 47 ◦ C and a critical time of 60 s [1]. Drill wear has the greatest influence on heat generation during the bone drilling process. In addition, the higher cutting forces and tool vibration caused by tool wear can cause damage or break the drill. The influence of increased temperature during the drilling process can be analyzed using a thermovision camera. Pre-operative planning of surgical, orthopedic, or trauma procedures, as well as practical training of doctors, are basic tools for shortening surgery times. Pre-operative training can be performed on non-fixed preparations, and has become the most advanced form of medical education. The preparation is almost identical to living tissue, apart from a lack of blood supply, however, availability is limited and approval from a Bioethics Committee is required. Additive techniques can also be used to produce anatomical bone models with various properties. With this approach, doctor have the opportunity to not only revise their anatomical knowledge but to also practice the selected surgical technique many times to avoid the enormous stress and sense of responsibility that accompanies performing a procedure on a living patient for the first time. Bone models are most often made of polymeric materials such as polyamide (PA, nylon), polyethylene terephthalate (PET), polylactide (PLA), acrylonitrile butadiene styrene (ABS) or polystyrene (PS), and high impact polystyrene (HIPS). Three-dimensional (3D) printing technology is the future of innovative medicine. Rapid prototyping (RP) of bone phantoms using 3D printing has a number of advantages including availability and cost of materials, easy storage of models, as well as repeatability and reproducibility. The models are based on computed tomography data and can imitate complex bone structures, diseases, or degeneration [10]. Fused deposition modeling (FDM) is one of the most popular RP methods, which involves melting plastic filaments then building each layer of the bone model one by one. As the layers cool, the material fuses to create the 3D object [4]. When creating models with this method, the following printing parameters should be taken into account: filling density, layer thickness, direction of application of subsequent layers, material type. Furthermore, the mechanical and thermal properties of the parent material are important. In this paper, the influence of 3D printing materials on bone phantom features are analyzed.

2 Materials and Methods In this study, simplified animal bone models were constructed using the following polymeric materials: PET with glycol, PLA (PLA Pro), and nylon (PA). Bone phantoms were prepared as cylinders with square infill patterns of various densities. The

Impact of 3D Printing Materials on Bone Phantom Features

181

infill ratio was 20, 40 and 60%, that means the sample with 60% of infill was nearest to a solid. Nevertheless, as the real bone is characterized with a sponge structure inside and a compact structure in the outer bone layer, there is no need to produce phantoms that are solid blocks. The sample with a 60% infill ratio was closest to a solid. Since real bone is characterized by an interior spongy layer surrounded by a more compact outer layer, there was no need to produce a solid phantom. The mechanical characteristics of the materials (based on literature data [6, 7]) were investigated using the Shore D hardness scale (elongation at break and hardness). The measured elongation at break was 4.3% for PLA, 5.8% for PET and 9.6% for PA. The measured hardness was 79.8 for PLA, 71.4 for PET and 83.0 for PA. The material properties are as follows: PLA plus is a polymer with a glass transition of 54–58 ◦ C, and a melting temperature of 205–230 ◦ C; PET has a glass transition temperature of 55–65 ◦ C, a melting temperature 165–180 ◦ C [3, 11]. Nylon has a melting temperature in the range of 190–350 ◦ C, but the material is typically melted at between 220–250 ◦ C for 3D pronting applications. Holes were drilled in nine samples, including three samples of each material in each of the defined infill ratios. A high-speed steel (HSS) drill bit with a diameter of 3 mm was selected. The drills were replaced after each material group was processed. Five holes were made in each sample with a 10mm offset. A cordless drill was used for the drilling process. Depending on the infill ratio of the sample, one hole required 5 to 12 s to be produced with a break of about 15 s between drilling successive holes. Rotation speed wasn’t recorded in the tests. It was fixed at a constant level by the manufacturer of the device. The specimens and the drills were observed using a stereoscopic microscope (SteREO Discovery.v8, Zeiss) with a 3x–80x digital zoom. Temperature changes in both the tested material and the tools were analyzed through thermal imaging studies. The temperature analysis was carried out using an infrared camera (FLIR A655sc) equipped with a 640 × 480 pixel microbolometer with a sensitivity of less than60 years), which are the most common surgically treated patient group, it is recommended to use validated femur replicates of osteopenic or osteoporotic type, to produce more representative and credible outcomes [2]. The purpose of the present study was to identify the apparent flexural modulus of cortical and trabecular bone analogue material applied in production of composite femurs of osteoporotic type by the SYNBONE company. This was done to obtain the material properties of composite femur components that can be later incorporated to numerical analysis of the same artificial femurs implanted with various orthopedic techniques. This was the first step necessary for further validation of the computer simulations by laboratory experiments.

2 Materials and Methods 2.1 Beam Samples The Synbone synthetic left femur bone of low density of both trabecular and cortical parts was tested (length: 450 mm, head diameter: 48 mm, canal diameter: 12 mm; Synbone Article Number LD2386). Lower density of the bone structure is made to resemble the osteoporotic tissues. The samples for 3-point bending test were prepared in the following manner: (a) beam samples prepared only from trabecular tissue of low density, (b) beam samples prepared from cortical tissue of low density, (c) composite samples prepared as beam composed of trabecular and cortical layer (Fig. 1). The beam samples were cut out of the femur shaft trying to get the most rectangular shape as possible. Six femurs were used for preparation of specimens. It

Flexural Modulus ...

235

Fig. 1 Beam cortical-trabecular composite samples

Table 1 Geometric properties of trabecular beam samples Sample no.

Width [mm]

Height [mm]

1

10.16 ± 0.26

6.10 ± 0.25

2

8.97 ± 0.18

5.79 ± 0.66

3

9.08 ± 0.65

3.87 ± 0.25

4

8.00 ± 0.54

5

7.15 ± 0.78

Length [mm]

Mass [g]

Apparent density [g/cm3 ]

84.00

1.14

0.22

93.00

1.07

0.22

96.00

0.80

0.24

5.27 ± 0.09

96.00

1.06

0.26

4.49 ± 0.40

95.00

0.94

0.31

Table 2 Geometric properties of cortical beam samples Sample no.

Width [mm]

Height [mm]

Length [mm]

Mass [g]

Apparent density [g/cm3 ]

1

11.72 ± 0.59

3.58 ± 0.30

92.00

1.64

0.43

2

11.55 ± 0.33

3.33 ± 0.12

96.00

1.74

0.47

3

10.81 ± 0.54

3.76 ± 0.33

94.00

1.97

0.52

4

11.46 ± 0.53

3.88 ± 0.31

96.00

2.11

0.49

5

14.26 ± 0.45

3.43 ± 0.16

86.00

2.02

0.48

6

9.39 ± 0.53

3.32 ± 0.38

96.00

1.33

0.44

was possible to obtain three samples from one femur. Next, the beam samples were measured by electronic caliper (LIMIT, Sweden, accuracy 0.02 mm) at both edges and in the middle of specimens. Finally, each specimen was weighted (LIMIT, Sweden, Iem-7, accuracy 0.01 g) and the apparent density was calculated as mass/volume relation, where volume was obtained based on measurements and upon assumption that samples were cuboid shaped. The mean values of specimens particular dimensions were used for these calculation. The obtained values of width b, height h, length L, mass and apparent density for trabecular, cortical and composite beam specimens are reported in Tables 1, 2 and 3, respectively. Standard deviation (denoted as ±) was also reported where appropriate.

236

K. Zerdzicki

Table 3 Geometric properties of composite (trabecular+cortical) beam samples Sample no.

Width [mm]

Height TRAB [mm]

Height COR [mm]

Height TOTAL [mm]

Length [mm]

1

16.94 ± 0.20

3.38 ± 1.45

3.88 ± 0.77

7.27 ± 1.48

196

2

16.51 ± 0.59

3.63 ± 1.26

3.84 ± 0.78

7.47 ± 1.55

194

3

18.14 ± 1.21

4.28 ± 1.37

3.87 ± 0.74

8.16 ± 1.74

196

4

16.37 ± 0.46

4.05 ± 1.31

3.92 ± 0.75

7.96 ± 1.61

194

5

19.06 ± 1.30

4.00 ± 1.63

4.29 ± 0.87

8.29 ± 1.91

192

2.2 Testing Procedure The universal testing machine Zwick/Roell Z0/20 (Zwick GmbH & Co. KG, Ulm, Germany) was used to apply the forces for three point bending tests through dedicated machine grips. Deflection of each sample was measured with accuracy of 0.01 mm by the videoextensometer (Zwick GmbH & Co. KG, Ulm, Germany) that followed black markers put in the middle of each sample and on the fixed base of the tasting machine. It is recommended by ISO 178 standard to form the specimens for bending tests that the distance between supports should be at least sixteen time greater than height of the specimen. It was impossible to get identical height for three types of beams tested, therefore for trabecular and cortical beams the support distance was L = 76 mm, while for composite beams it was L = 130 mm. The initial load for single beams was 0.5 N, while for composite beams it was 2 N, selected arbitrary after some preliminary tests. For all the tests the load was force controlled with the rate of 10 mm/min. The tests lasted till the significant drop of force level for cortical beams (Fig. 2(a)) or the fracture of trabecular beams (Fig. 2(b)). It was observed that trabecular beams fractured in the middle, while the cortical beams revealed highly plastic behavior and deforms permanently without sign of fracture.

2.3 Elastic Flexural Modulus Identification Basic transformation of the equation of the deflection in the middle of a beam under three point bending loading gives the flexural modulus in the following form: F=

P L3 · δ 48I

(1)

Flexural Modulus ...

237 20

16

Trabecular beam no. 1 Trabecular beam no. 2 Trabecular beam no. 3 Trabecular beam no. 4 Trabecular beam no. 5

14

15

Force [N]

Force [N]

12 10 8

10

6 Cortical beam no. 1 Cortical beam no. 2 Cortical beam no. 3 Cortical beam no. 4 Cortical beam no. 5 Cortical beam no. 6

4 2

5

0

0 0

5

10

15

20

25

0

30

2

4

6

8

10

12

Deflection [mm]

Deflection [mm]

(b)

(a)

Fig. 2 Force-displacement curves for beams made of polyurethane material analogue to cortical tissue (a), trabecular tissue (b) Table 4 Flexural modulus for cortical analogue material Sample no.

IC O R [mm4 ]

P/δ [N/mm]

FC O R [MPa]

1

44.674

0.986

201.85

2

35.531

0.807

207.71

3

47.759

1.239

237.26

4

55.926

1.531

250.36

5

47.825

1.615

308.83

6

28.539

0.564

180.74

Average

43.376 ± 9.800

1.124 ± 0.410

231.12 ± 45.58

where P is the force acting in the middle of the sample resulting in deflection δ, L is support span length, and I = bh 3 /12 stands for moment of inertia of the cross section of the beam tested. The support span length L and moment of inertia I are dependent on the sample’s geometry, while the force P and deflection δ are obtained from experimental tests. The relation P/δ corresponds to the slope of the linear part of the force-displacement curve and was identified in the deflection range of 0–7 and 0–5 mm for cortical and trabecular tissue analogue, respectively. The approximation of that slope by linear function was performed in SigmaPlot commercial software. Every specimen of cortical and trabecular type was analyzed separately. The parameters from Tables 1, 2 were used to calculate the moment of inertia I and next, the flexural modulus F for cortical (IC O R , FC O R ) and trabecular (IT R AB , FT R AB ) beams (Tables 4 and 5, respectively).

238

K. Zerdzicki

Table 5 Flexural modulus for trabecular analogue material Sample no.

I T R AB [mm4 ]

P/δ [N/mm]

FT R AB [MPa]

1

191.93

2.608

124.27

2

145.15

2.212

139.37

3

43.99

0.682

141.80

4

97.39

2.664

250.16

5

54.03

1.881

318.39

Average

106.50 ± 62.32

2.009 ± 0.807

194.80 ± 8.54

Fig. 3 Force-deflection curves for composite cortical-trabecular beams

40

Force [N]

30

20

Composite beam no. 1 Composite beam no. 2 Composite beam no. 3 Composite beam no. 4 Composite beam no. 5

10

0 0

5

10

15

20

Deflection [mm]

3 Results and Validation For validation of the identified values of flexural modulus, the three point bending tests were conducted on composite beam samples (Fig. 3). The deflection of beam samples were calculated in the elastic region of behavior according to the transformed relation 1 and with the reduction of the geometric properties of the trabecular material part, to get composite beam characterized by the cortical analogue material only. It was done for three force levels: 10, 15, and 20 N taking into account the samples dimensions (Table 3) and identified flexural modulus (Tables 4, 5). The obtained values were then compared with experimental deflections and the relative error was calculated and reported as well (Table 6). The good convergence between theoretical and experimental results is observed. The sparse differences are probably caused by the simplified protocol of measuring the beam samples, which were not ideally rectangular and regular in shape.

Flexural Modulus ...

239

Table 6 Deflection of composite samples - theoretical versus experimental Sample no. 1

2

3

4

5

Force [N]

Experimental [mm]

Theoretical [mm]

Error [%]

10

2.67

3.99

32.96

15

4.37

5.98

26.90

20

6.19

7.98

22.41

10

3.21

3.76

14.82

15

5.41

5.65

4.25

20

7.81

7.53

3.74

10

2.67

2.63

1.43

15

4.43

3.95

12.08

20

6.29

5.27

19.34

10

2.74

3.14

12.51

15

4.54

4.70

−3.36

20

6.54

6.27

4.25

10

2.11

2.38

11.66

15

3.49

3.58

−2.28

20

4.95

4.77

3.91

4 Discussion In the last few years more orthopedic interest is put on the studies of non-healthy tissues, osteoporotic and osteotonic, as fracture occur within patients of lower bone condition [8]. Consequently, these patients most often requires fixation, where the implant is positioned in weak bone structure. In one of the well-known study on tissue retrieved from human femoral head [11], the apparent density, compressive modulus and compressive stiffness of osteoporotic trabecular was found 0.38 g/cm3 , 2.5 MPa (range 0.6–5.8 MPa) and 247 MPa (range 50–410 MPa), respectively, comparing to healthy normal tissue of 0.47 g/cm3 , 3.3 MPa (range 0.4–9.0 MPa) and 310 MPa (range 40–460 MPa), respectively. More up to date research [8] delivered results, where the apparent Young’s modulus of cortico-cancellous mixed bone blocks retrieved from vertebrae, obtained by inverse FEM analysis was found to be 374 MPa ± 208 (range 87–791 MPa). These values should be reproduced by polyurethane (PU) foams used for production of synthetic bones dedicated for implant testing. In several studies mechanical properties of the commercial foams used as bone analogs were investigated concerning various densities, foaming direction and inner foam structure [4, 12, 15, 16]. Study of the PU foam supplier for orthopeadic device evaluation (General Plastics (GP), Tacoma, WA, USA) found that compressive elastic modulus and strength values ranges from 115 to 794 MPa and 4.7 to 24.7 MPa, respectively, for foams having densities of 0.240–0.641 g/cm3 [4]. The compressive Young’s modulus and yield strength of the Sawbones PU foam mimicking trabecular cancellous tissue for the density of 0.16 g/cm3 and 0.32 g/cm3 was found to be in the range 15.1–151.4 MPa and 0.9–4.5 MPa, respectively [14]. Concerning different types of foaming, named cellular, solid and open, the Young’s

240

K. Zerdzicki

modulus values for Sawbone PU foams were from 55.5 ± 8.27 MPa to 505 ± 9.70 MPa, from 50.5 ± 0.83 MPa to 461 ± 3.55 MPa and from 36.8 ± 6.14 to 260 ± 66.8 MPa, respectively, depending on the density level [12]. In comparison, for production of bone surrogates, SynboneTM uses foams of density 0.08–0.67 g/cm3 according to ISO 1183 standard that corresponds to range 5–42 PCF of ASTM 1622 standard. Available types of foams are characterized by different compressive and flexural strength, that are identified according to relevant standards (ASTM-D 695, ASTMD 790, ISO 604, ISO 178) and reported officially. However, there is no detailed information which foam is used for production of osteoporotic type of bone. For the Synbone surrogate bone analyzed within this study, the only information states that the femur is characterized by cortical low density and soft cancellous bone. Therefore, it is of the greatest importance to experimentally examine the cortical and trabecular analogue material withdrawn from particular composite femur type before performing corresponding finite element modelling of bone/implant constructs. Furthermore, experiments are usually performed in foaming direction, which corresponds to bone analog application [4], but can thus influence the final results, what was reported by [12], who got Young’s modulus average values about 23–31% different depending on the testing and foaming direction. Additionally, both leader producers of bone surrogates apply different techniques for manufacturing their final artificial bone and use components of differ densities and mechanical properties, depending of product application. Consequently, it is crucial to investigate material finally used in the bone replicate that was done within this study and e.g. research of [5], who studied trabecular osteoporotic tissue blocks retrieved directly from the natural and synthetic femoral head in opposite to other studies, where experiments on synthetic foams are conducted on the foam blocks provided by the producers. The blocks were cut out of the femoral head in the plane perpendicular to implanting screw positioning. The study found that compressive stiffness of the blocks (dimensions about 13 × 13 × 25 mm) was 863.2 ± 509.56 N/mm (range 107.5–2141.6 N/mm) and the density 1 ± 0.06 g/cm3 (range 0.9–1.1 g/cm3 for the natural trabecular osteoporotic bone, compared to the stiffness of 215.9 ± 57.8 N/mm (range 127–306 N/mm) with the density 2.67 g/cm3 ± 0.4 (range 2.1–3.2 g/cm3 ) for synthetic Synbone cancellous analog material. The data of preliminary research presented in this study can serve as a reference to help design, model and analyze tests using replicate femur bones of Synbone producer. Acknowledgements This study was financed by grant supporting young researchers at Faculty of Civil and Environmental Engineering Gdansk University of Technology in year 2020/2021. Daria Albertynska is acknowledged for assistance during experiments and help in initial results overworking.

Flexural Modulus ...

241

References 1. Basso T, Klaksvik J, Syversen U, Foss OA (2012) Biomechanical femoral neck fracture experiments-a narrative review. Injury 43(10):1633–1639 2. Basso T, Klaksvik J, Syversen U, Foss OA (2014) A biomechanical comparison of composite femurs and cadaver femurs used in experiments on operated hip fractures. J Biomech 47(16):3898–3902 3. Brown AD, Walters J, Zhang Y, Saadatfar M, Escobedo-Diaz JP, Hazell PJ (2019) The mechanical response of commercially available bone simulants for quasi-static and dynamic loading. J Mech Behav Biomed Mater 90:404–416 4. Calvert KL, Trumble KP, Webster TJ, Kirkpatrick LA (2010) Characterization of commercial rigid polyurethane foams used as bone analogs for implant testing. J Mater Sci Mater Med 21(5):1453–1461 5. Ceynowa M, Zerdzicki K, Klosowski P, Pankowski R, Roclawski M, Mazurek T (2020) Drill holes decrease cancellous bone strength: a comparative study of 33 paired osteoporotic human and 9 paired artificial bone samples. PloS One 15(10):e0241143 6. Ceynowa M, Zerdzicki K, Klosowski P, Pankowski R, Rocławski M, Mazurek T (2020) The early failure of the gamma nail and the dynamic hip screw in femurs with a wide medullary canal. a biomechanical study of intertrochanteric fractures. Clin Biomech 71:201–207 7. Cristofolini L, Viceconti M (2000) Mechanical validation of whole bone composite tibia models. J Biomech 33(3):279–288 8. El Masri F, Sapin de Brosses E, Rhissassi K, Skalli W, Mitton D (2012) Apparent young’s modulus of vertebral cortico-cancellous bone specimens. Comput Methods Biomech Biomed Eng 15(1):23–28 9. Heiner AD (2008) Structural properties of fourth-generation composite femurs and tibias. J Biomech 41(15):3282–3284 10. Heiner AD, Brown TD (2001) Structural properties of a new design of composite replicate femurs and tibias. J Biomech 34(6):773–781 11. Li B, Aspden RM (1997) Composition and mechanical properties of cancellous bone from the femoral head of patients with osteoporosis or osteoarthritis. J Bone Miner Res 12(4):641–651 12. Marter AD, Dickinson AS, Pierron F, Fong YK, Browne M (2019) Characterising the compressive anisotropic properties of analogue bone using optical strain measurement. Proc Inst Mech Eng Part H J Eng Med 233(9):954–960 13. Papini M, Zdero R, Schemitsch E, Zalzal P (2007) The biomechanics of human femurs in axial and torsional loading: comparison of finite element analysis, human cadaveric femurs, and synthetic femurs 14. Patel PS, Shepherd DE, Hukins DW (2008) Compressive properties of commercially available polyurethane foams as mechanical models for osteoporotic human cancellous bone. BMC Musculoskelet Disord 9(1):1–7 15. Shim V, Boheme J, Josten C, Anderson I (2021) Use of polyurethane foam in orthopaedic biomechanical experimentation and simulation. Polyurethane 171–200 16. Thompson MS, McCarthy ID, Lidgren L, Ryd L (2003) Compressive and shear properties of commercially available polyurethane foams. J Biomech Eng 125(5):732–734

Informatics and Modelling in Biomedical Engineering

The use of mathematical modeling in the analysis of phenomena occurring in the human body is an opportunity to learn and describe processes in which the use of direct measurement methods is impossible or unethical. Advanced mathematical calculations allow, among other things, to accurately estimate the value of forces and stresses in selected anatomical structures, determine how blood flows through the heart, lungs and the blood vessels or indicate the kinematic dependences describing individual motor tasks. The topics of the works presented in the chapter include numerical determination of mechanical properties of head protection systems, comparative analysis of models for predicting immunogenicity of viral antigens and analysis using the Finite Element Method of stress distribution of implants. In addition, the work also included research on blood flows or research on the impact of reading speed on the understanding of the text.

The Application for Reading Comprehension and Reading Speed Test Łukasz Grabny, Rafał Doniec, Szymon Siecinski, ´ Natalia Piaseczna, and Konrad Duraj

Abstract Vision is considered one of the most important human senses. People acquire a large amount of data by reading, so the faster someone reads, the more information he/she should learn. However, this is not always the case. Electrooculography (EOG) plays a significant role in research on reading. We recorded the EOG signals with the JINS MEME ES_R smart glasses from eight subjects who read a text and then completed a reading comprehension test implemented as a multiple-choice test. Reading speed was calculated as the time of reading the text with a constant number of words. The visualization of reading was based on detecting the saccades in the horizontal component of the EOG signal. The algorithm consists of signal preprocessing and FCM (fuzzy c-means) thresholding. Three types of saccades and their duration were detected and associated with blinking, eye movement to the left, and eye movement to the right. The developed application shows the reading process with the use of an animation of eye-walking on the text. The application tests carried out show that it works properly. Keywords Electrooculography · Wearable devices · Reading speed · Reading comprehension

Ł. Grabny Faculty of Biomedical Engineering, Silesian University, Roosevelta 40, 41-800 Zabrze, Poland e-mail: [email protected] R. Doniec · S. Sieci´nski (B) · N. Piaseczna · K. Duraj Faculty of Biomedical Engineering, Department of Biosensors and Processing of Biomedical Signals, Silesian University, Roosevelta 40, 41-800 Zabrze, Poland e-mail: [email protected] R. Doniec e-mail: [email protected] N. Piaseczna e-mail: [email protected] K. Duraj e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_25

245

246

Ł. Grabny et al.

1 Introduction Vision is the most important human sense [7] and helps obtain information about the surroundings [23]. The people acquire most data by reading [13]. Therefore, we can state that the faster a person reads, the more information can learn. However, this is not always the case, because the reading comprehension is a complex process, which involves the language phonology and grammar rules [22]. Human eyes move constantly because only the fovea provides a small area of high resolution vision needed in seeing various objects. The most important eye movements for reading are saccades and fixations [4, 22]. Saccade is a fast movement of the eyes, reaching 500 °/s [1], and even 1000 °/s, which last 20 to 200 ms [5]. The duration of the saccade depends on the angular distance which is linear up to 30° [6]. During the saccadic movement the eye is unable to register the image, therefore they harmonize with the leading movements [16]. Saccades can be forced or spontaneous and may be followed by corrective saccades which are the movements which occur when the eye misses the desired target [3]. The eye movements in reading can be characterized as a sequence of repeated small saccades responsible for the eye movements between the words, separated by large, regressive saccades which occur while going to the next line [4]. Many methods have been designed to measure eye movement, such as oculography. One of them is electrooculography (EOG) which measures the electric potential between the cornea and the retina of the human eye. This potential generates an electric field that changes its orientation as the eye moves [12]. The most important eye movements for reading are saccades and fixations [22]. The EOG plays a significant role in research of reading. Eye movements and eyesight analysis can be used to detect reading [15], measure its parameters, and also help in learning to read to people with impaired vision [24]. Along with the development of technology, many new solutions have emerged in eye tracking, such as smart glasses, which may be used to test the concentration, sleepiness or reading activities in a non-invasive way [8]. The average reading speed of an adult is 180–250 words/minute [2]. However, there are techniques that can increase the reading speed. Some people are able to read up to 1000 words but it comes at the cost of understanding the text (omitting some words). These techniques include [22]: – elimination of phonetization, i.e. pronouncing the words read in your mind and its special form—articulation (movements of lips), – skimming—quickly movements of eyes across the text in search of specific words and key information, – expanding the peripheral vision of the eyes to reach more words in the eyesight (from a few words to a few lines). However, its limitation is the visual acuity of the fovea, – shortening the fixation time. Measurement of reading speed is a simple task performed by counting the number of words of a given text and measuring the time of reading. On the other hand,

The Application for Reading Comprehension and Reading Speed Test

247

measurement the effectiveness of reading (reading comprehension), is usually performed as a multiple-choice test that refers to the previously read text [22]. To the best of our knowledge only Kunze et al. performed the study on reading which involved the use of JINS MEME smart glasses [14]. The purpose of our study to provide a method for recording the reading speed and reading comprehension with smart glasses.

2 Material and Methods 2.1 Experiment Setup EOG signals were recorded using the JINS MEME ES_R smart glasses a 3-axis accelerometer, 3-axis gyroscope and three dry EOG electrodes made of stainless steel [10, 13]. The smart glasses are equipped with a Bluetooth module to provide wireless communication with a computer or a mobile device. To establish the connection between the computer and the start glasses via Bluetooth, we connected a USB dongle with a Bluetooth module and used the ES_R Development Kit application available at [11]. The sampling frequency of the EOG signal was 100 Hz [10].

2.2 Signal Processing There are many methods for detecting saccades [19]; In our approach the detection of saccades consists of signal preprocessing, classification and encoding presented as the flowchart in Fig. 1. The preprocessing phase consists of resampling, baseline drift removal and noise attenuation. The classification phase consists of fuzzy c-means clustering and determining the detection threshold. Due to the uneven signal sampling related to Bluetooth transmission, the signal was resampled by creating a new time vector with a constant sampling rate 100 Hz. The signal was interpolated using the interp1 function in MATLAB

Fig. 1 Saccades detection algorithm flowchart

248

Ł. Grabny et al.

(Natick, MA, USA). The next step was removing the baseline drift which was estimated by modeling as a polynomial [21]. In our case the baseline drift in horizontal and vertical EOG signals was modeled by a 20th degree polynomial by using the polyfit MATLAB function. The values of the estimated baseline drift were calculated using the polyval function. Then, the estimated baseline was subtracted from the signal. The 20th degree of polynomial was determined experimentally to ensure the optimal fit to the baseline of the EOG signal. The noise in the EOG signal may be caused by incorrect electrode placement, muscle activity or power line interference. Aforementioned artifacts affect significantly the quality of the signal, which hampers the recognition of eye movements, especially saccades. To attenuate the noise we applied discrete wavelet transform (DWT). Discrete wavelet transform decomposes the signal into an approximation component and details component. The approximation component is implemented as the low-pass filter and downsampling by two. The details component is obtained by performing high-pass filtering and decimation [9]. The wavelet function is defined as:   1 t −b (1) ψab (t) = √ ψ a a where ψ is the mother wavelet, a is the scaling parameter, b is translation parameter. To attenuate the noise we used the symlet5 wavelet, the scaling based on the noise level estimate based on the first-level coefficients, the fourth level of decomposition and fixed-form threshold for wavelet denoising defined as: th =

√ 2 ln N

(2)

where N is the number of signal samples. The classification phase consists of fuzzy c-means (FCM) clustering and determining the detection threshold. Due to the inter-subject variation, we detect the saccades by using the thresholds determined fuzzy c-mean clustering. Our implementation is based on the approach of Pander et al. [20]. The FCM clustering is based on dividing the data set of N elements into c groups of similar objects. The fuzzy membership assumes that an element may belong to more than one group. The membership value u determines the membership of an element to a given set and is between 0 and 1. The group prototype vi is defined as: N ∀1≤i≤c vi = k=1 N

(u i j )m x j

k=1 (u i j )

m

(3)

where x j is the feature vector representing j-th element, m denotes the fuzziness of groups, u i j ∈ [0, 1] is an i-th element of u matrix (the membership of i-th element to the j-th group) and wi = (∀i = 1, 2, 3, . . . , c) are the weights of elements in a group.

The Application for Reading Comprehension and Reading Speed Test

249

Because we can distinguish three types of saccades, we set the number of groups to three, the number of iterations was limited to 20 and the minimum value of membership function value u > 0.75. The membership function matrix is defined as: 1 (4) ui j = 2 c  ||x j −vi ||  m−1 k=1

||x j −vk ||

where m > 1. The FCM algorithm aims to minimize the function: Jm =

N  C 

u imj ||x j − vi ||2 , 1 ≤ m ≤ ∞

(5)

j=1 i=1

The saccades were detected as the local maxima and minima were detected with findpeaks MATLAB function. The saccades below the first threshold were classified as the first type and the saccades above were the saccades of the second type. Local minima were detected for the following parameters: MinPeakProminence (minimal amplitude of the peak [17]) set as p = 0.7|th|, where th is the threshold value in μV, and MinPeakDistance (minimal distance between peaks [17]) was 90 samples. To detect local maxima, the MinPeakProminence value was set to 1.5 µV in order to exclude very small peaks related to the possible residual noise. The results are shown in Fig. 2. Detected saccades were encoded as follows: zero (0) was assigned to eye movements to the left and one (1) to eye movements to the right.

Fig. 2 Detected saccades. An asterisk (*) denotes the eye movement to the right, a triangle denotes the eye movement to the left

250

Ł. Grabny et al.

Reading Speed Measurement. In our study we measured the reading speed by measuring the time needed for reading a predefined text with a known number of words. The text, unknown to the readers, contains information about the biography of Nikola Tesla and consists of 228 words. The time of reading the text was measured as the time between the start of recording and the end of recording. Equation (6) defines the reading speed in words per minute (words/min) as: reading speed [words/min] =

number of words × 60 reading time [s]

(6)

2.3 Software We used MATLAB to design signal processing algorithms. The graphical user interface was implemented in .NET Windows Presentation Foundation (WPF) because of the capability to create the animation of reading a text. Graphical User Interface is presented on the Fig. 3. The reading speed test starts after clicking “Start” button which launches JINS MEME Academic Logger, the software used to acquire EOG signal. The next step was to choose the COM port of the smart glasses and the device identification number. Then, the text with a constant number of words (228) was loaded into the TextBlock control (WPF widget for text with formatting and effects [18]). The measurement of the reading speed is based on measuring the time (see the Eq. (6)) after clicking “Choose a signal” button.

Fig. 3 Graphical User Interface of implemented application; Reading text animation process

The Application for Reading Comprehension and Reading Speed Test

251

Reading Comprehension Test. The reading comprehension test was implemented as a simple multiple-choice test with single-select questions. The test consisted of ten questions about the text used to measure reading speed. After completing the test, the subject is ask to type the name and then, the result is shown as the percentage of correct answers along with the correct answers.

3 Results 3.1 Performance Measures of the Algorithm The outcome of saccade detection is considered positive if a saccade is present in the signal or negative if no saccade is detected. Consequently, the method of detecting saccades can be characterized by: – true positive detections (TP)—number of correctly detected saccades, – false positive detections (FP)—number of falsely detected saccades (if the detector finds a saccade in a place where there is no real saccade), – false negative detections (FN)—number of undetected saccades, – true negative detections (TN)—number of signal changes that the detector correctly classified as a non-saccade. The performance of measurements were expressed as accuracy (Acc), sensitivity (Se), specificity (Sp), and positive predictive value (PPV) defined as in Eqs. 7, 8, 9, and 10: Se =

TP T P + FN

(7)

Sp =

TN FP + T N

(8)

P PV =

Acc =

TP T P + FP

TP +TN T P + T N + FP + FN

(9)

(10)

The performance measures of saccade detection are presented in Table 1. In the Table 1 we consider the case of two classes—saccade detection and no saccade occurence (binary classification). The classifier achieved high sensitivity (over 85%), specificity (over 85%), accuracy (over 86%) and PPV (over 95%) in all subjects. The highest sensitivity was achieved for the subject 1 (96.24%) and the lowest sensitivity was observed for the subject 6. The lowest specificity, PPV

252

Ł. Grabny et al.

Table 1 Performance measures of saccade detection Subject TP FP TN FN Se [%] 1 2 3 4 5 6 7 8 Total

179 190 63 94 132 136 72 150 1016

9 7 1 1 2 1 1 1 23

68 101 33 60 84 125 51 143 665

7 9 5 14 17 40 12 21 125

96.24 95.48 92.65 87.04 88.59 77.27 85.71 87.72 89.04

Sp [%]

PPV [%] Acc [%]

88.31 93.52 97.06 98.36 97.67 99.20 98.08 99.30 96.66

95.21 96.45 98.44 98.95 98.51 99.27 98.63 99.34 97.79

93.92 94.79 94.12 91.12 91.91 86.42 90.44 93.01 91.91

and accuracy were observed for the subject 1, subject 1, and subject 6, whereas the highest specificity, PPV and accuracy were observed for the subject 1, subject 8, and subject 2. This set of parameters confirms that the applied classifier is reliable.

3.2 Application Performance In order to test the application, eight people were asked to read the text as quickly as possible and take the reading comprehension test. The results of the reading comprehension tests and measurements of reading speed are shown in Table 2. Table 2 Reading speed in words per minute (words/min) and the results of the reading comprehension test Subject number

Reading speed [words/min]

Reading comprehension test [%]

1

155.07

80

2

165.20

80

3

169.14

50

4

189.13

40

5

181.12

60

6

365.58

50

7

201.48

60

8

223.76

70

Mean

206.31

61.25

SD

63.56

13.64

The Application for Reading Comprehension and Reading Speed Test

253

The mean reading speed was 206.31 words/min but the most reported measurements were between 150–175 words/min.

4 Conclusion and Discussion In our study we developed an application for measuring reading speed and testing reading comprehension by using smart glasses, which can visualize the process on reading by tracking the eyes with the EOG. The developed detection algorithm used preprocessing and fuzzy clusterization (namely fuzzy c-means). The application tests confirmed its reliability. Unfortunately, visualization of reading process was not fully implemented due to the problems caused by the third party application used to record the EOG signals. The reason is probably the need of pairing with smart glasses. An additional burden is the large selection of measurement options, which strongly affects the user interface. Another limitation was also the inability to test the subjects who wear the glasses. The proposed detection method is highly efficient and precisely determines the temporal position of saccadic eye movements, even with a low signal-to-noise ratio. The presented method may have potential application in EOG signal processing systems for the determination of visual acuity. The visualization of reading was determined by the quality of the recorded EOG signal, affected by blinking and the contact of skin and electrodes. The animation of reading the text was reliable in 7 out of 9 cases (77% of cases). Due to the small number of participants the study should be considered a pilot study. However, the developed application works properly as long as the EOG signal has no artifacts. Eye movement analysis is valuable in both clinical and scientific work. An application with animation for tracking text and then assessing the level of understanding can be useful for research on the level of awareness in learning new content. The applied classifier with the accuracy of circa 92% allows the application to be used in more difficult clinical cases related to rehabilitation after vision loss or visual impairment. In future studies we will consider adding more texts to be used during the tests of reading speed and comprehension and implementing signal acquisition without using third-party applications provided by the manufacturers of smart glasses. Another issue to be addressed is to include smaller saccades in visualization of reading and consider also other methods for detecting saccades.

References 1. Buczkowska H (2016) Porównanie wybranych parametrów ruchów sakadowych oczu oraz okre´slonych etapów funkcji czytania w grupie dzieci z ambliopia˛ i bez ambliopii [The comparison of selected parameters of saccades and specific states of reading function on children with amblyopia and without ambpyopia]. Ph.D. thesis, Medical University of Pozna´n, Poland

254

Ł. Grabny et al.

2. Buzan T(2003) Podre˛cznik szybkiego czytania [The Speed Reading Book]. Ravi 3. Cafasso A, Karlsson S (2017) Automatic detection of saccadic eye movements using EOG for analysing effects of cognitive distraction during driving. Master’s thesis, Chalmers University of Technology, Göteborg, Sweden 4. Chandra J, Krügel A, Engbert R (2020) Modulation of oculomotor control during reading of mirrored and inverted texts. Sci Rep 10(1):4210 5. Fischer B, Ramsperger E (1984) Human express saccades: extremely short reaction times of goal directed eye movements. Exp Brain Res 57(1):191–195. https://doi.org/10.1007/ BF00231145 6. Huda K, Hossain MS, Ahmad M (2015) Recognition of reading activity from the saccadic samples of electrooculography data. In: Conference: 2015 international conference on electrical and electronic engineering (ICEEE), vol 1, November 2015 7. Hutmacher F (2019) Why is there so much more research on vision than on any other sensory modality? Front Psychol 10:2256 8. Ishimaru S, Kunze K, Tanaka K, Uema Y, Kise K, Inami M (2015) Smart eyewear for interaction and activity recognition. In: Proceedings of the 33rd annual ACM conference extended abstracts on human factors in computing systems - CHI EA 2015, ACM Press 9. Jedli´nski Ł (2012) Odszumianie danych rejestrowanych wielokanałowo z u˙zyciem transformaty falkowej. Eksploatacja i Niezawodno´sc´ 14(2):145–149 10. JINS Inc. (2020) JINS, Inc. researchers. Available at: https://jins-meme.com/en/products/es/. Accessed 19 Nov 2020 11. JINS Inc. (2020) JINS MEME documentations. https://jins-meme.github.io/. Accessed 19 Nov 2020 12. Joseph DP, Miller SS (1991) Apical and basal membrane ion transport mechanisms in bovine retinal pigment epithelium. J Physiol 435(1):439–463 13. Kanoh S, Ichi-Nohe S, Shioya S, Inoue K, Kawashima R (2015) Development of an eyewear to measure eye and body movements. In: 2015 37th annual international conference of the IEEE engineering in medicine and biology society (EMBC). IEEE, August 2015 14. Kunze K, Katsutoshi M, Uema Y, Inami M (2015) How much do you read? In: Proceedings of the 6th augmented human international conference on - AH 2015, ACM Press 15. Latifoglu F, Ileri R, Demirci E, Altintop CG (2020) Detection of reading movement from EOG signals. In: 2020 IEEE international symposium on medical measurements and applications (MeMeA), vol 1, November 2020 16. Leigh R, Zee D (2015) The neurology of eye movements. Contemporary neurology series. Oxford University Press 17. MathWorks Inc. (2020) Find local maxima - MATLAB findpeaks. MATLAB Documentation. https://www.mathworks.com/help/signal/ref/findpeaks.html. Accessed 19 Nov 2020 18. Microsoft Corporation: Textblock class. Microsoft Docs. Available at: https://docs.microsoft. com/en-us/dotnet/api/system.windows.controls.textblock?view=net-5.0 Accessed 19 November 2020 19. Pander T et al (2012) A new method of saccadic eye movement detection for optokinetic nystagmus analysis. In: 2012 annual international conference of the IEEE engineering in medicine and biology society, pp 3464–3467, August 2012 20. Pander T, Czaba´nski R, Wróbel J, Horoba K, Przybyła T (2013) An application of fuzzy clustering for saccade detection in ENG signal. Aktualne Problemy Biomechaniki 7:137–142 21. Pettersson K, Jagadeesan S, Lukander K, Henelius A, Hæggström E, Müller K (2013) Algorithm for automatic analysis of electro-oculographic data. BioMed Eng OnLine 12(1):110 22. Rayner K, Schotter ER, Masson MEJ, Potter MC, Treiman R (2016) So much to read, so little time. Psychol Sci Public Interest 17(1):4–34 23. Schifferstein HN (2006) The perceived importance of sensory modalities in product usage: a study of self-reports. Acta Psychologica 121(1):41–64 24. Sibert JL, Gokturk M, Lavine RA (2000) The reading assistant: eye gaze triggered auditory prompting for reading remediation. In: Proceedings of the 13th annual ACM symposium on user interface software and technology. UIST 2000, pp 101–107. Association for Computing Machinery, New York

Numerical Investigations of Mechanical Properties of Head Protection Systems Against the Effects of Dynamic Loads Aleksandra Je˛drzejewska, Kamila Wi´sniewska, Monika Ratajczak, and Tomasz Klekiel

Abstract The aim of the study is to analyse head protection systems against dynamic loads. The fully protecting construction of helmet against the destruction of the tissues of the head resulting from loads has not been developed yet. Therefore, researches in this issue are needed. The study undertook the development and numerical research of materials absorbing loads energy for military use. The work analyzes two absorbing constructions with conical and oval holes for five different Young’s modulus of construction materials (2, 10, 20, 30 and 200 MPa). The research allowed to determine the possibility of use of the proposed absorber structures for head protection. The research was carried out using finite elements method (FEM) and have shown that the best energy attenuation is provided by the absorber with oval holes and the Young’s modulus does not significantly affect its properties. In the case of conical-hole absorber, an influence of the Young’s modulus on its protective properties was observed. Keywords Energy absorbers · Numerical simulation · FEM · Head injuries · Strength analysis

A. Je˛drzejewska · K. Wi´sniewska · M. Ratajczak · T. Klekiel Department of Biomedical Engineering, Faculty of Mechanical Engineering, University of Zielona Góra, 65-516 Zielona Góra, Poland e-mail: [email protected] M. Ratajczak e-mail: [email protected] T. Klekiel e-mail: [email protected] A. Je˛drzejewska (B) Science and Technology Park of University of Zielona Góra, 66-002 Zielona Góra, Poland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_26

255

256

A. Je˛drzejewska et al.

1 Introduction The brain is one of the most important organs in the human body and protecting it from any injuries is a priority. The Centers for Disease Control and Prevention (CDC) defines Traumatic Brain Injury (TBI) as “disruption of the normal functioning of the brain that may be caused by a blow, impact or shock to the head, or penetrating head trauma”. Depending on the severity of the brain injuries, a person with TBI may experience changes in the level of consciousness that can range from bewilderment and confusion to loss of consciousness [25]. Memory loss may also occur. Traumatic Brain Injury is usually caused by a sudden collision of the head against a stationary obstacle - a sudden acceleration/deceleration in the head area that causes the brain to collide back and forth with the skull structure due to an impact [21]. These injuries can happen during road accidents, especially vehicle-to-bicyclist accidents, during which the cyclist often does not have a helmet, which may result in a significant impact of the dynamic loading on the structures of the skull and the sutures between them [18]. However, it is worth nothing that they are also the result of military conditions [20]. Traumatic brain injury (TBI) has been called a “signature injury” of Iraq and Afghanistan Conflicts. The Defense and Veterans Brain Injury Center (DVBIC) report nearly 350,000 incident diagnoses of TBI in the U.S. military since 2000 [13]. It is also worth nothing that potentially non-hazardous objects may contribute to injuries in these conditions. Burkacki et al. showed that hitting both light and heavy rifles can lead to serious head injuries - during an explosion, their free movement within the cabin of a military vehicle can turn them into deadly objects [2]. Numerous statistical cases of head injuries indicate problems with head protection systems. So far, no solution has been found that would fully protect the head tissue from degradation as a result of an accident. Therefore, the work is aimed at developing better head protection solutions than before. As technology advances, preventing these injuries may become more effective [1, 28] – It reduces the deceleration of the skull, and hence the brain movement, by managing the impact. The soft material incorporated in the helmet absorbs some of the impact and therefore the head comes to a halt more slowly. This means that the brain does not hit the skull with such great force. – It spreads the forces of the impact over a greater surface area so that they are not concentrated on particular areas of the skull. – It prevents direct contact between the skull and the impacting object by acting as a mechanical barrier between the head and the object [27]. – Achieving these functions is possible thanks to the cooperation of its four main components, which are the shell, the energy-absorbing lining, the comfort lining, as well as straps and fasteners [27] shown in Fig. 1. The aim of the study was to improve the protective systems of the head in order to better protect the brain tissues against permanent damage, which is motivated by the already mentioned fact of the prevalence of these injuries and insufficiently

Numerical Investigations of Mechanical Properties of Head Protection Systems ...

257

Fig. 1 Components of the helmet: 1- shell, 2 - energy absorbing-lining, 3 - comfort lining, 4 - straps and fasteners; own elaboration based on [9, 27]

optimized helmet structures, taking into account, for example, as demonstrated in the work of Jamroziak et al., based on the analysis of the trend of currently used military helmets in the largest armies of the world (USA), as well as the Polish army, it was concluded that these helmets are able to withstand 2 out of 3 threats, the impact of which is smaller, but not the most destructive, namely an explosion [10]. In order to conduct the research, numerical investigations were used. Mathematical modeling using finite element method (FEM) has many significant advantages, such as the low cost of numerical experiments with the ability to change basic parameters, such as material properties, impact velocity and it allows us to gain a large range of results that are difficult to measure non-invasively with equipment, including stress-strain or force magnitudes [12]. Considering this, the described method was used to determine the absorbing properties of the designed energy-absorbing linings and the effect of the impact energy on the white and gray matter of the brain. The research focused on to determine the influence of selected shapes of absorbers on the protection of brain tissues, however, it is worth nothing that the tests were not carried out only for the absorber layer, because its properties can be determined only on the basis of the analysis of stresses and deformations of the head structures. The model also takes into account the elements of the head, which makes it possible to compare the ability to absorb the impact energy, and thus the protection of the head, by the proposed geometries of the absorers. Two shapes, absorber layer with conical holes and oval holes, were selected for the study. The absorber layer with conical holes is a solution that works well for helmets bikes to create entire linings instead of a few loose elements [5]. As for the oval-hole absorber, this geometry is a modification of the geometry containing the conical holes, also this is layer with holes, but in the shape of ovals. Therefore, the influence of the shape of the holes will be checked on the effect of energy absorption.

258

A. Je˛drzejewska et al.

2 Materials and Methods 2.1 Model Development and Materials The developed models consist of 9 layers, among which: shell, absorber, skull, cerebrospinal fluid, white and gray matter. The first two layers are elements of the helmet, and the rest of the heads. The model size is 145x135 mm. The material parameters for all model layers are presented in Table 1. Several different Young’s modulus were assigned for the absorber layer, as the study investigated the impact of both the absorber geometry and properties of construction material on stresses and deformations of the head tissues. The value of Young’s modulus was: 2, 10, 20, 30 and 200 MPa and was selected taking into account the fact that in the literature the most frequently used absorber material is EVA, the stiffness of which is 20 MPa. The selected range of values contained greater and lesser values than for EVA to determine the effect of this parameter on the absorbing properties. Due to the fact that the subject of the research was the absorber, the simplest possible (linear isotropic) model of the shell material (Kevlar) was used. This allowed for simplification of calculations and determination of the absorber properties. The model layers have been bent to reflect the natural curvature of the head, with the exception of the fixed layer, which is the point of impact. The layers lie on top of each other. Two models different in the geometry of the helmet’s absorbing layer were analysed. The first model was shown in Fig. 2A, in this model the absorber layer has an conical holes. The second model with oval holes in the absorber layer was shown in Fig. 2B. The models consist of head and helmet fragments to reduce the calculation time. The tests were not carried out only for the absorber layer, because its properties can be determined only on the basis of the analysis of stresses and deformations of the head structures. The model also takes into account the elements of the head, which makes it possible to compare the ability to absorb the impact energy, and thus the protection of the head, by the proposed geometries of the absorers.

Fig. 2 The section of simulation model of the head and helmet with an absorber layer with: A conical holes and B oval holes, composed of the following layers: 1-white matter, 2-gray matter, 3-cerebro-spinal fluid, 4-pia mater, 5-dura mater, 6-skull, 7-tested absorber, 8-shell and 9-stationary obstacle

Numerical Investigations of Mechanical Properties of Head Protection Systems ...

259

Table 1 The parameters of each layer of models Material

Layer of model

Type of material

White matter

White matter

Viscoelastic

Parameters*

Layer thickness (mm)

ρ=1060 [26]

25

E=4 kPa [26] pr =0.45 [26] Momentary shear modulus: G0=1040 Pa [24] Viscoelastic constant: 80 s−1 [24] Gray matter

Gray matter

Viscoelastic

ρ=1060 [26]

2.5 [5]

E=4 kPa [26] pr =0.45 [26] Momentary shear modulus: G0=680 Pa [24] Viscoelastic constant: 80 s−1 [24] Cerebro-spinal fluid

Cerebro-spinal fluid

Linear isotropic

ρ=1000 [26]

Pia mater

Pia mater

Linear isotropic

ρ=1130 [3]

Linear isotropic

ρ=1140 [26]

2 [15]

E=0.15 MPa [26] pr =0.49999 [26] 7 [6]

E=15 MPa [3] pr =0.45 [3]

Dura mater

Dura mater

1.1 [7]

E=31,5 MPa [26] pr =0.45 [26]

Compact bone

Skull

Linear isotropic

ρ=2000

Linear isotropic

ρ=100

Linear isotropic

ρ=1440 [19]

Linear isotropic

ρ=7820 [8]

7 [16]

E=15000 MPa [26] pr =0.22 [26]

Absorber material

Absorber

Kevlar

Shell

30 [26]

E=2, 10, 20, 30, 200 MPa pr =0.45 10 [17]

E=58 GPa [19] pr =0.3 [22]

Steel 40H

Stationary obstacle

E=207 GPa [8] pr =0.3 [8]

*ρ - density

[kg/m3 ],

E - Young’s modulus, pr - Poisson’s ratio

10

260

A. Je˛drzejewska et al.

2.2 Simulation Conditions The values of stress, energy and deformation during dynamic loading on the structures of the brain were analysed. The simulation was performed using the initial speed of 6 m/s given to the elements of the helmet and the head towards the fixed obstacle. This speed was chosen on the basis of a literature analysis of studies of protective helmets used by soldiers [4, 23]. The simulation conditions took into account the effect of the force of gravity (against the direction of speed).

3 Results and analysis The stress distribution for the absorber layer absorber layer with conical holes and oval holes as well as 2 and 200 MPa of Young’s modulus is shown in Fig. 3. Smaller changes in the stresses value on the skull with the change in the Young’s modulus of the material were observed for the absorber layer with oval holes than with conical holes. The absorber layer with oval holes is less susceptible to changes in the stiffness of the material. In this model was observed also a smaller stress distribution. The absorber layer absorbs the energy formed on the helmet shell and thanks to which the head structures layers are experienced on less stress than the shell. It confirms the good absorption properties of proposed structures. The greatest stress, below the absorber layer, occurred in the skull. For the absorber with oval holes, the highest value of Young’s modulus of 200 MPa (Fig. 3B) resulted in higher stresses than

Fig. 3 Stress distribution for the absorber layer with: oval holes (A and B) and conical holes (C and D) using Young’s modulus 2 MPa (A and C) and 200 MPa (B and D)

Numerical Investigations of Mechanical Properties of Head Protection Systems ...

261

Young’s modulus of 2 MPa (Fig. 3A). For the absorber with oval holes, the lowest value of Young’s modulus of 2 MPa (Fig. 3C) resulted in higher stresses than Young’s modulus of 200 MPa (Fig. 3D). Therefore, the properties of energy absorbers in helmets depend both on the shape of the holes and structural properties of the material. Figure 4 shows the distribution of strain for the absorber with oval and conical holes for Young’s modulus of material 2 and 200 MPa. For the conical holes of absorber layers, there were greater strain of head tissues than for oval holes absorber. The absorber with oval holes better protects the head tissues against strain. This shape of the holes of the absorber is also less susceptible to changes in the Young’s modulus of the material. For the oval shape of the absorber holes, smaller strains were observed for 200 MPa than for 2 MPa of Young’s modulus. The inverse relationship applies to the conical shape of the absorber holes, for 2 MPa the strain is smaller than for 200 MPa of Young’s modulus. The conical holes of absorber at 2 MPa of Young’s modulus were deformed the most. For both shapes of the absorber holes, a dependence of the degree of protection against strain on the value of the Young’s modulus of the material was observed. The numerical analysis of stress (Fig. 5) and strain (Fig. 6 and 7) of skull and brain structures allowed to determine protective properties of the proposed shapes of absorber holes of helmet protecting head. The brain and the skull are the most vulnerable to injuries, therefore an analysis was performed for these structures. The presented figures show the values of stresses and strains of the skull and brain as a result of an impact for various absorber holes shape and Young’s modulus of material. A greater dependence of the stress value on Young’s modulus was observed for the tapered holes of the absorber, and the stress value is greater for these holes than for the oval holes for both the skull and the brain (Fig. 5). The absorption properties

Fig. 4 Distribution of strain for the absorber layer with: oval holes (A and B) and conical holes (C and D) for Young’s modulus 2 MPa (A and C) and 200 MPa (B and D)

262

A. Je˛drzejewska et al.

Fig. 5 Stress on skull and brain resulting from the impact simulation, depending on the holes geometry and Young’s modulus of the absorber layer

of absorber with oval holes are more stable and less dependent on the material properties. This analysed structure absorb the impact energy without transferring to the brain tissue and skull the stresses greater than the allowable, respectively 34 kPa and 153 MPa [24]. Thus, both hole shapes meet the strength criteria and protect both the brain and the skull. It has been observed that decreasing the Young’s modulus to a certain value reduces the stresses on the skull to a minimum, where further reduction of the Young’s modulus causes an increase in stress. For both holes, the minimum stress occurred at the Young’s modulus of 10 MPa for the skull. In the case of stresses on the brain for conical holes, two minima were observed at 30 and 2 MPa Young’s modulus, while no significant changes were observed for oval holes. Based on literature reports [14], it was established that the maximum permissible strain on the skull is 0,025 mm/mm. The strain and strain energy on the skull were shown on Fig. 6. Both absorber hole structures protected the skull against strain, smaller strain occurred for oval holes also strain energy. It has been observed that decreasing the Young’s modulus to a certain value reduces the strain and strain energy on the skull to a minimum, where further reduction of the Young’s modulus causes an increase in strain especially for conical-holes absorber. For conical-holes absorber the minimum value of strain was observed at 10 MPa of Young’s modulus for the skull, while no significant changes were observed for oval holes. It was possible to determine a better geometry of the absorber holes by analyzing the strain of the brain in Fig. 7. Based on literature reports [24], it was established that the maximum permissible strain on the brain tissues is 0,2 mm/mm. Only for oval holes, this value was not exceeded, so the absorber with oval holes has properties that allow the brain to be sufficiently protected. The absorber with conical holes requires

Numerical Investigations of Mechanical Properties of Head Protection Systems ...

263

Fig. 6 Strain and strain energy on skull resulting from the simulated impact, depending on the holes geometry and Young’s modulus of the absorber layer

Fig. 7 Strain and strain energy on brain resulting from the simulated impact, depending on the holes geometry and Young’s modulus of the absorber layer

modification of the structure, e.g. by changing the number of cones per unit, in order to improve its properties. Similarly to the stress analysis, it was found that the strain of the brain is changed by changing the Young’s modulus of the absorber material. The change is greater for conical holes. The oval-hole absorber therefore has better absorption properties. It has been observed that change of the Young’s modulus caused change of brain strain value. In the case of strain on the brain for conical holes, two minima were observed at 30 and 2 MPa Young’s modulus, for strain energy minima was observed for 20 and 2 MPa of Young’s modulus. No significant changes were observed for oval holes.

264

A. Je˛drzejewska et al.

4 Conclusions Taking into account all the configurations, the best energy attenuation is provided by the absorber with oval holes and the Young’s modulus does not significantly affect its properties. In the case of conical-hole absorber, an influence of the Young’s modulus on its protective properties was observed. In the case of the absorber with conical holes, an influence of the Young’s modulus of material on its protective properties was observed. This structure has the best absorption properties for 2 and 30 MPa of Young’s modulus taking into account the stresses and strains of the brain, and 10 MPa taking into account the strains of the skull. The Young’s modulus value does not significantly affect the properties of the oval-holes absorber. An absorber with an oval hole better protects the brain and the deformations inside tissues are about 12% less than for protection by the layer of conical hole.absorber. Both proposed structures show good mechanical properties allowing the absorption of the impact energy, however, in the case of the conical-holes geometry, the strain of the brain is higher and this is insufficient to protect from damage. This geometry should be structurally modified to improve its ability to strain protection. Studies by Jazi et al. [11] show that with an increase in stiffness, stresses and strains on the tissues may increase, which was confirmed in this research with a maximum increase in stiffness up to 200 MPa.

References 1. Arkusz K, Klekiel T, Sławi´nski G, Be˛dzi´nski R (2019) Influence of energy absorbers on Malgaigne fracture mechanism in lumbar-pelvic system under vertical impact load. Comput Methods Biomech Biomed Eng, pp 1–11 2. Burkacki M, Wola´nski W, Sucho´n S, Joszko K, Gzik-Zroska B, Sybilski K, Gzik M (2020) Finite element head model for the crew injury assessment in a light armoured vehicle. Acta Bioeng Biomech 22:173–183, https://doi.org/10.37190/ABB-01556-2020-02 3. Depreitere B (2004) A rational approach to pedal cyclist head protection, PhD dissertation. K.U, Leuven 4. De-Shin L, Yao-Te Ch (2017) A finite element investigation into the impact performance of an open-face motorcycle helmet with ventilation slots. Appl Sci 7:279 5. Galvanetto U, Caserta G, Blanco D, Ghajari M, Cernicchi A (2013) Helmet performance and design. Helmet Research in the WP3 of the MYMOSA Project. Proceedings of the 1st International Conference on Helmet Performance and Design, London, UK HPD-2013-9 6. Hazel B (2014) Impact mechanics of helmet components, University of Twente, RMIT University, 8–17 7. Howay L (2010) Characterization and modification of helmet padding system to improve shockwave dissipation. Clemson University, TigerPrints 8. https://www.intechopen.com/books/textiles-for-advanced-applications/contemporarypersonal-ballistic-protection-pbp (13.03.2021) 9. https://velmet.ua/en/49-sredstva-zasshity/shlema-i-kaski/ballisticheskiy-shlem-tor-d-skaverom-coyote.html (28.05.2021)

Numerical Investigations of Mechanical Properties of Head Protection Systems ...

265

10. Jamroziak K, Bajkowski M, Bocian M, Polak S, Magier M, Kosobudzki M, Stepien R (2019) Ballistic head protection in the light of injury criteria in the case of the Wz.93 combat helmet. Appl Sci 9:2702 11. Jazi M, Rezaei A, Ghodrat K, Ziejewski M (2014) A computional study of influence of helmet padding materials on the human brain under ballistic impacts. ResearchGate 17:1368–1382. https://doi.org/10.1080/10255842.2012.748755 12. Levadnyi I, Awrejcewicz J, Zhang Y et al (2018) Finite element analysis of impact for helmeted and non-helmeted head. J Med Biol Eng 38:587–595 13. Lindquist LK, Love HC, Elbogen EB (2017) Traumatic brain injury in Iraq and Afghanistan veterans: new results from a national random sample study. J Neuropsychiatry Clin Neurosci 3254–259. https://doi.org/10.1176/appi.neuropsych.16050100 14. McElhaney JH, Fogle JL, Melvin JW, Haynes RR, Roberts VL, Alem NM (1970) Mechanical properties of cranial bone. J Biomechanics 3:495–511. https://doi.org/10.1016/00219290(70)90059-x 15. Mills N, Gilchrist A (2006) Bicycle helmet design. IMechE, vol 220 Part L: J Materials: Design and Applications 16. National Center for Biotechnology Information. http://www.ncbi.nlm.nih.gov (26.05.2021) 17. National Research Council (2014) Review of department of defense test protocols for combat helmets. Washington, DC: The National Academies Press. https://doi.org/10.17226/18621 18. Ptak M, Ratajczak M, Kwiatkowski A, Sawicki M, Wilhelm J, Fernandes FAO, Druszcz A (2018) Investigation of biomechanics of skull structures damages caused by dynamic loads. Acta Bioeng Biomech 4:143–150, https://doi.org/10.5277/ABB-01252-2018-03 19. Quinn JA (2002) Composites—Design Manual. 3, James Quinn Associates Ltd 20. Ratajczak M, Klekiel T, Sławi´nski G, Be˛dzi´nski R (2020) Investigation of helmet-head interaction in the aspect of craniocerebral tissue protection. Curr Trends Biomed Eng Bioimages Anal: Proc 21st Pol Conf Biocybern Biomed Eng, 2019 21. Ratajczak M, Ptak M, Chybowski L, Gawdzi´ska K, Be˛dzi´nski R (2019) Material and structural modeling aspects of brain tissue deformation under dynamic loads. Materials (Basel) 12:1–14 22. Salam SS, Mehat NM, Kamaruddin S (2019) Optimization of laminated composites characteristics via integration of Chamis Equation, Taguchi method and Principal Component Analysis. Jt Conf Green Eng Technol Appl Comput 2019, IOP Conf Ser: Mater Sci Eng 551, https://doi. org/10.1088/1757-899X/551/1/012110 23. Sandberg M, Tse K, Long B, Lee Hp (2018) A computational study of the EN 1078 impact test for bicycle helmets using a realistic subject-specific finite element head model. Comput Methods Biomech Biomed Eng 21:1-9. https://doi.org/10.1080/10255842.2018.1511775 24. Tse KM, Lim SP, Tan VBC, Lee P (2014) A review of head injury and finite element head models. Am J Eng, Technol Soc 1:28-52 25. U.S. Department of Veterans Affairs https://www.research.va.gov/topics/tbi.cfm (28.05.2021) 26. Wang F, Lee HP, Lu C, Cheng QH (2005) Evaluation of human head injury in tracked vehicle subjected to mine blast w Michael D. Gilchrist (ed). IUTAM Proc Impact Biomech: Fundam Insights Appl, pp 273-280, Springer 27. WHO, https://www.who.int/roadsafety/projects/manuals/helmet_manual/1-Why.pdf (28.05.2021) 28. Wi´sniewska K, Je˛drzejewska A, Ratajczak M, Klekiel T (2021, May 19–21) Analysis of the mechanical properties of impact absorbing structures used in military helmets. Biocybern Biomed Eng-Curr Trends Callenges, Proc 22nd Pol Conf Biocybern Biomed Eng, Warsaw, Poland - in print

Influence of the Reorientation Function on Brodmann Areas Detection Efficiency Ilona Karpiel

Abstract Preprocessing and analysis of medical data are very important. Information technology has become the right hand of doctors and without there medicine would not be the same. Great attention should be paid to the possibilities offered by computer science as well as to the functions that we can use. Functions, parameters, and algorithms used strongly influence the final result of each analysis. The analysis of both fMRI and rsfMRI should be preceded by the so-called "manual corregistration". This function allows the data to be compared later, as the xyz point is set exactly at the AC location. The work shows how the analysis process can run and what influence the preliminary alignment has on the analysis. It should be noted that it would be useful to systematize the so-called pre-processing. … Keywords Signal processing · SPM · Data analysis · Brain

1 Introduction The human brain consists of about 52 regions, numbered by the German anatomist Korbinian Brodmann in the early 1900’s. To this day, it is the most interesting part of the human body and despite a very large amount of research, it is not yet 100% known. Modernity has given the opportunity to "look" deep into the human brain. Unfortunately, the research methods themselves, such as magnetic resonance imaging without the use of specialized methods of analysis, do not allow to present e.g. brain functionality. Matlab software (MathWorks, Inc.) and SPM (Statistical Parametric Mapping) toolbox are most often used for this purpose. The basic analysis of functional research consists of the so-called reorientation, realignment, normalizaI. Karpiel (B) Łukasiewicz Research Network, Institute of Medical Technology and Equipment, Zabrze, Poland e-mail: [email protected] URL: http://www.itam.lukasiewicz.gov.pl © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_27

267

268

I. Karpiel

tion, and smooth. The so-called manual corregistration consists of positioning the cursor, i.e. the position 0 0 0 at the so-called AC (anterior commissure) point. AC is a transversely oriented commissural fiber tract connecting the left and right cerebral hemispheres. It is a very important anatomical landmark that connects different parts of the limbic system on both sides and plays a role in the interhemispheric transfer of visual, auditory, and olfactory information between temporal lobes, however, the role of the AC is not yet well understood. The AC corresponds to a white matter tract almost surrounded by gray matter that crosses the midline just ventral to the supraoptic recess of the third ventricle and the columns of the fornix. It ends in the amygdaloid nucleus of the temporal pole bilaterally. Jean Talairach proposed a “three-dimensional proportional grid which is based on a set of anatomical landmarks: AC and posterior commissure (PC), the midline sagittal plane, and the exterior boundaries of the brain at each edge and proposed the use of the anterior and posterior commissures as reference points for the brain coordinate system. Given these landmarks, the origin (zero-point) in the three-dimensional space is defined as the point where the AC intersects the midline sagittal plane. The axial plane is then defined as the plane along the AC/PC line that is orthogonal to the midline sagittal plane, and the coronal plane is defined as the plane that is orthogonal to the sagittal and axial planes [24]. AC with the PC interconnects the two cerebral hemispheres of the brain and also interconnects the amygdalae and temporal lobes, contributing to the role of memory, emotion, speech and hearing. It also is involved in olfaction, instinct, and sexual behavior. Talairach continued to work even after his retirement in 1978 until the age of 96 years.25 and to this day the system he proposed is in use [23, 27–29]. The coordinate systems provide a link between the physical structures in the brain and the coordinates in the image. We need a common space in which different individuals can be aligned. The MNI templates the most common templates used for spatial normalization are those developed at the Montreal Neurological Institute, known as the MNI templates. MNI space is proposed as an international brain coordinate system. A series of images must be realigned, which is briefly called a motion correction. While data acquisition, patients usually move slightly their head, which means: the placement of spatial images may vary within the same coordinate system (see [4, 14, 21, 25]). The goal of motion correction is to adjust the time series of images so that the brain is in the same position in every image (see [16] and [6]). Different people have brains that are different sizes, and even when the same person is scanned multiple times, the brain will be in different places in the image depending upon exactly where the head was positioned in the scanner [24]. The purpose of this paper is to show the effect of the operation of the first pre-processing step on the fMRI study results especially patients pre and postoperative and to prove that a small change in the set of initial conditions (location of the AC point) can lead to a massive difference in the outcome.

Influence of the Reorientation Function ...

269

2 Material and Methods 2.1 Subjects and FMRI Data Acquisition Three preoperative patients with different lesions and three postoperative were included in the study, comprising 3 females and 3 males. Scanning was performed at the Helimed Diagnostic Imaging Center in Katowice as part of normal clinical work using a 1.5T MAGENTOM Aera scanner (Siemens, Erlangen, Germany), equipped with a 20-channel head-neck coil. Additionally used a computer with a projector to allow display of earlier prepared paradigms (such as finger tapping) in PsychoPy software. The protocol included 1) Ultrafast Gradient Echo 3D sequence (3D T1-MPRAGE, TE/TR= 2.67 ms/1900 ms, TI=1100 ms, slice thickness 1.0 mm, resolution 1 × 1 × 1 mm3 , FOV=250 mm × 250 mm, matrix size=256× 256, 192 axial slices; 2) fMRI images (EPI SE sequence, resolution 3 × 3 × 3 mm3 , TE/TR = 50 ms/3140, TI = 1100 ms, slice thickness 4.0 mm, FOV=1320 mm x 1320 mm, matrix size=384 × 384, 36 axial slices per volume, covering both hemispheres).

2.2 Data Analysis Analysis of brain activations based on EPI SE sequence was performed in the statistical parametric mapping (SPM 12 package, http://www.fil.ion.ucl.ac.uk/spm) in MATLAB (MathWorks, Inc.) environment. Before data analysis, has been made converting DICOM (Digital Imaging and Communications in Medicine) images to NIfTI (Neuroimaging Informatics Technology Initiative) format [7]. The next step was ‘manual coregistration which was the point 0 setting in Anterior Commissure (AC) (Fig. 1). The commissure has become key landmarks because Talairach chose them to determine the standard alignment of the brain for their famous atlas. To align a scan Talairach space, a brain has to be set so that the anterior commissure (AC) and posterior commissure (PC) are on a horizontal line. It is relatively easy to see on most structural scans. Ref. point is set on anatomical and additionally applied for all functional images. It is essential and necessary for future steps. Have been selected all functional images and one image structural (file with the new point). After "reorientation images" all images both structural and functional have set points in the same position. You can check it once again by using “check reg”. The function can be used to check if two or more scans are matched onto each other. This check should be performed after each step of spatial ’manual’ pre-processing. The time for manual correction (first step) can take from 5-10 minutes/one patient depending. Later step by step analysis of one patient takes about 20 minutes. The analysis of each patient occurred twice and was divided into 2 stages. It consisted of: analysis without manual correction (use of pre and postprocessing) and the analysis with the activity of setting the point x,y,z (0,0,0) in point AC (using pre and post-processing) to compare the results of these two analyses (Fig. 2). The initial (i.e., "wrong") position in our exper-

270

I. Karpiel

iment was never acceptable (Fig. 2ab), and the final position has been shown in Figs. 1 and 2c. There is no way for the other researcher to consider the i.e. presented in Fig. 2ab items as correct for further preprocessing. Generally in medical teams, data processing/analysis (pre and post-processing) is done by medical physicists (Radiologist gets finished results). The user has ensured that the analysis is repeatable. The AC point setting (as in Fig. 1) was repeated 20 times and no change in the results was noticed. The comparison is not presented in this publication because it did not contribute relevant information. This suggests the "repeatability" of the experiment at the preprocessing stage, but no other scientist/clinician, or physician was consulted at this stage. The steps of the data analysis process consisted of spatial pre-processing: ‘realignment’, ‘coregistration’, and ‘spatial smoothing’. In these analyzes, we used a model based on the Tstatistic, where selected significance level (p) defined probability. The automatic analysis was not performed in the work presented. No scripts were used.

Fig. 1 Anterior commisure in presented in 3 planes: Coronal, Sagittal, Axial, zoom 80 × 80

Influence of the Reorientation Function ...

271

Fig. 2 Position point before a,b and after c the manual coregistration structural image

3 Results and Discussion The author of this work focused on one of the basic steps that are done before using more advanced pre-processing methods. The subject of comparing results from many different experiments, either on the same person at different times, or between different people, raises many issues, only a few of which are covered here. Setting the AC point is a basic element at the beginning of pre-processing, which needs to be taken into account to avoid the so-called false results in the future. Even the smallest change can have a significant impact on results. When the data is loaded, the 0,0,0 point is very arbitrary, depending on the camera, patient setup, etc. This is why the "manual corregistration" step is used for standardization, where each analysis would start at the AC point. A pixel located 2-3 mm on may already have a different location and this is what the analysis presented here is about. The process itself takes about 10 minutes along with loading the data. Generally, the analysis is performed by experienced physicists, but there is nothing to prevent someone else from doing it. Tables 1 and 2 show the highest values for a person with and without a set AC point. As can be seen from the tables, the different cursor positions have resulted in different results. The T-value for patient number 1 was 13.09 and later 10.54. For each patient, the T-value has changed, even though the analysis and the preprocessing itself have not changed. What’s more, the highest values that were located using the Talairach Client’s software were located in other Brodmann areas based on slices from the brain of a single person. It uses a proportional grid system based on the location of two anatomical landmarks, the Anterior Commissure (AC) and the Posterior Commissure (PC). For patient number 1, the highest value was located in the right and left hemispheres. This shows that it is not entirely correct to describe the methodology in different types of publications and to compare to published works. Figure 3 shows the analysis of the same patient without a pre-set AC point and after setting the Ac point and normalizing. Standardized area (coordinates: 48, -30, 0) represents the Wernicke area (b), while in the picture "a" - cross is seen in a completely different place in the brain. In Fig. 4 the situation is similar, which only confirms the validity of the manual corregistration function. The same coordinates are in different places. It should be noted that with such discrepancies,

272

I. Karpiel

Table 1 The highest activation value for each preoperative patient (without a set AC and with AC) ID patients

Volume voxels T-value

1

33890 33436

2 3

Coordinates (xyz) mm

Brodmann area

13.09

13,−41,−44

Right Cerebellum, BA30

10.54

−1,−62,−44

Left Cerebellum, Posterior Lobe

30802

09.46

−26,−52,−15

Left Cerebellum, Occipital Lobe

30253

9.06

−27,−45,−15

Left Cerebellum, Anterior Lobe

36073

11.37

30,23,95

None

35631

09.37

−27,−45,−15

None

results cannot be compared without prior knowledge of the methodology. It is known that normalization is necessary to compare results. However use of masking as well as normalization in patients with lesions is still controversial and is the subject of much work lately [2, 8]. Unfortunately, not everyone mentions the mentioned AC point. Along with setting the starting point, the patient’s head position can also be corrected. Sometimes the image and their position is "asymmetrical" and can be manually corrected, and example values for this operation are shown in Table 3. The changes are little visible but is necessary. Some images may need a larger value therefore manipulate your scan until you are satisfied with the result (note: rotations are given in radials; 180 degrees is 3.1416 radials). These are two very important points of pre-processing analysis. It should also be mentioned that the presented results are using 6mm kernel, and many scientists also use the remaining values (28mm) and various statistical methods described in detail by a group of scientists in the 1990s (see [1, 3, 9–13, 15, 17, 20, 22, 26, 30]) which again shows how difficult it is to compare the results. Some activations cannot be observed using the selected statistics, the number of p, the number of voxels, or a certain number of software or toolboxes based on some other algorithms. One of the most interesting and extensive publications is the aforementioned publication " Variability in the analysis of a single neuroimaging dataset by many teams" [5], which shows that the topic is still attractive and important. Therefore, the author of the publication entitled "Influence of the reorientation function on Brodmann Areas detection efficiency" decided to refer to the basic but elementary process of pre-processing, which is manual corregistration, which aims to raise awareness among "young scientists" and more. Table 2 additionally shows the impact of standardization on data analysis and the location of selected areas. The analyses showed that the voxel volume changes 10 times. T-value for each analysis is different but not as much as shown in table no.1. It should be mentioned here that as far as we know where the cursor should be, it is still a "manual" setting of the beginning of x y z by the user. Each subsequent user performing the same analysis process can perform it a little differently. The author presented two different, similar positions of the AC point in point 000 after the analysis was repeated. This is very dangerous in terms of results. Usually, one analysis chosen by the user is performed and not several variants are tested. These is

Influence of the Reorientation Function ...

273

Fig. 3 Postoperative patients: Images of activations on Wernicke significance level p 0.001 for extent threshold 0 voxels. a point 0 no setting in AC, b position AC set in structural and functional images. Significance level p 0.001 for extent threshold 0 voxels. Analysis with normalization

the initial users of various software or, as in this case, the SPM toolbox and MATLAB software. The author presents only "two" examples for each patient because only ONE point is right (AC) and every other point is wrong. The standardization and normalization of this selected step are based on the selection of only the AC point and that’s what the user’s aiming for (Figs. 1 and 2c). In 2017, DPARSF software appeared [6], which was used by users. Very friendly, intuitive, and saves time. The authors presented a solution that automatically allows setting the AC point in the right place. SPM allows users to set it manually, however, not every researcher has carried out the so-called manual correction before the analysis. The method itself has already been described in the manual to SPM, but the subject of analysis, methods of analysis, pre and post-processing steps return with every year. The topic is still not exhausted, and one can even say highly controversial because of the freedom of choice of all the components mentioned above. The publication entitled: "Variability in the analysis of a single neuroimaging dataset by many teams" from 2020 [5] presents data analyses from all over the world (researchers analyzed the same set of data). This is a very interesting work showing how you can have a different approach and how many results you can get based on the same set of data. Other aspects related to the use of parameters such as the kernel and "p" have already been described in previous papers [18, 19]. The inter/intraobserver analysis is available. Additionally, the analysis is repeatable. The analysis for one patient was performed 20 times (manually) to see if the result was the same. The same result was obtained. However, this applies to initial manual corregistration only (first step). Each next step is also possible, but this applies to knowing 100% of the methodology. If the analysis was performed by several professionals not familiar with the details of pre- and post-processing, the results might be different (e.g. different t-value). Therefore, analysis in SPM has both advantages and disadvantages. Advantages, because you can freely conduct analysis, analyzing one patient with taking into account AC point or exceptionally not taking into

274

I. Karpiel

account (if, for example, you do not use additional programs such as Talairach Client / you do not want to compare results / there are no major shifts), normalizing patient, or for example, conducting group analysis. Is it possible to manually misalign the AC point ? If not guided by anatomy, you probably can. But the question remains, how much to believe the algorithms used in various types of software? There is also some kind of error/error limit. Automatic tools fully facilitate and relieve the user. However, in certain cases, it is good to perform some steps manually. The worst possible option during the analysis process is just not paying attention to the occurring shift of the central point (xyz coordinates) to the so-called center (AC). In summary, each successive step of pre and post-processing builds on the previous one, hence the importance of manual corregistration itself.

Table 2 The highest activation value for each postoperative patient (without a set AC and with AC). Preprocessing with normalization ID patients 4

5

6

Volume voxels

T-value

Coordinates (xyz) mm

Brodmann area

32438

11.48

−18,−54,−46

Left Cerebellum, Posterior Lobe, Cerebellar Tonsil, Gray Mattercre

219334

11.65

−18,−60,−10

Left Cerebellum, Posterior Lobe, Declive, Gray Matter,; Left Cerebellum, Anterior Lobe, Culmen, Gray Matter; Left Cerebrum, Occipital Lobe

36548

5.98

−14,−32,35

Left Cerebrum, Limbic Lobe, Cingulate Gyrus, White Matter,

205774

7.26

−62,−42,0

Left Cerebrum, Temporal Lobe, Middle Temporal Gyrus, Gray Matter, Brodmann area 21

37512

11.36

37,−28,87

None

212582

10.83

40,−22,54

Right Cerebrum, Parietal Lobe, Postcentral Gyrus, Gray Matter, Brodmann area 3, BA4

Table 3 Values for 3 different subjects. Setting the correct head position Subjects1 Subjects2 Subjects3 Pitch [rad] Roll [rad] Yaw [rad]

0.01 0.001 0

0 0.01 0.02

0.01 0.02 0.01

Influence of the Reorientation Function ...

275

Fig. 4 Preoperative patients: Images of activations on Wernicke significance level p 0.001 for extent threshold 0 voxels. a point 0 no setting in AC, b position AC set in structural and functional images. Significance level p 0.001 for extent threshold 0 voxels. Analysis with normalization

4 Conclusion A significant influence on the localization of anatomy in the brain using interactive tools is the setting of the so-called AC point at point 000. Skipping the first step can suggesting the wrong localization of activity. Results showed that analysis without set AC could indicate activation in the left hemisphere area or after set AC point in the right hemisphere or no an indication of activation at the area which can lead to incorrect descriptions and diagnosis. Correct setting of the AC point minimizes the risk of indicating the wrong location by additional tools supporting and identifying coordinates. Acknowledgements The author thanks Helimed Diagnostic Imaging and dr Aldona Giec-Lorenz for the data provided.

276

I. Karpiel

References 1. Aguirre G, Zarahn E, D’Esposito M (1998) The inferential impact of global signal covariates in functional neuroimaging analyses. NeuroImage 8:302–306 2. Andersen SM, Rapcsak SZ, Beeson PM (2010) NeuroImage 53:78 3. Andersson J (1997) How to estimate global activity independent of changes in local activity. NeuroImage 6:237–244 4. Ashburner J, Henson R, et al (2012) SPM manual, welcome trust center for neuroimaging 5. Botvinik-Nezer R, Holzmeister F, Camerer CF, et al (2020, June) Variability in the analysis of a single neuroimaging dataset by many teams. Nature 582(7810):84–88 6. Chao-Gan Y, Yu-Feng Z (2010, May 14) DPARSF: a MATLAB toolbox for “pipeline” data analysis of resting-state fMRI, Front Syst Neurosci 7. Cox R, Ashburner J, Breman H, Fissell K, Haselgrove C, Holmes C, Lancaster J, Rex D, Smith S, Woodward J, Strother S (2004) A (sort of) new image data format standard: Nifti-1, in 10th Annual Meeting of the Organization for Human Brain Mapping (OHBM 2004), vol 25. Budapest, Hungary, June, pp 13–17 8. Crinion J, Ashburner J, Leff A, Brett M, Price C, Friston K (2007) NeuroImage 37:866 9. Descombes X, Kruggel F, von Cramon DY (1998) Spatio-temporal fMRI analysis using Markov random fields. IEEE Transactions on Medical Imaging 17(6):1028–1039 10. Eddy W, Fitzgerald M, Noll D (1996) Improved image registration by using Fourier interpolation. Magn Reson Med 36(6):923–931 11. Friston K, Jezzard P, Turner R (1994) Analysis of functional MRI time-series. Hum Brain Mapp 1:153–171 12. Friston K, Worsley K, Frackowiak R, Mazziotta J, Evans A (1994) Assessing the significance of focal active tions using their spatial extent. Hum Brain Mapp 1:214–220 13. Friston K, Holmes A, Poline J-B, Grasby P, Williams S, Frackowiak R, Turner R (1995) Analysis of FMRI time series revisited. NeuroImage 2:45–53 14. Friston KJ (2002) Brain mapping: the methods, 2nd edition, Chapter 22: Statistics I: Experimental Design and Statistical Parametric Mapping, Elsevier 15. Hu X, Le T, Parrish T, Erhard P (1995) Retrospective estimation and correction of physiological fluctutaion in functional MRI. Magn Reson Med 34(2):201–212 16. Huettel S, Song A, McCarthy G (2004) Functional magnetic resonance imaging. Sinauer Associates, Sunderland, MA 17. Jenkinson M, Smith S (2001) Global optimization for robust affine registration of brain images. Med Image Anal 5(2):143–156 18. Karpiel I, Klose U, Drzazga Z (2019) Optimization of rs-fMRI parameters in the Seed Correlation Analysis (SCA) in DPARSF toolbox: a preliminary study. JNeuro Res 97:433–443 19. Karpiel I, Drzazga Z, Mazgaj P, Drzazga M, Giec-Lorenz A (2018) Optimization of fMRI analysis of speech areas in pre- and postoperative diagnostics, Acta Physica Polonica A 133(3):s. 728–730 20. Kruggel F, von Cramon D, Descombes X (1999) Comparison of filtering methods of fMRI datasets. NeuroImage 10:530–543 21. Nisbet R, Elder J, Miner G (2009) Statistical analysis and data mining applications, Elsevier 22. Poline J-B, Mazoyer B (1994) Enhanced detection in brain activation maps using a multifiltering approach. J Cereb Blood Flow MetabIsm 14:639–642 23. Rahman M, Murad GJ, Mocco J (2009) Early history of the stereotactic apparatus in neurosurgery. Neurosurgical Focus 24. Russell A Poldrack, et al (2011) Handbook of functional MRI data analysis 25. Smith SM (2004) Overview of fMRI analysis. Br J Radiol 77:167–175 26. Schmitt F, Stehling M, Turner R (1998) Echo-planar imaging: theory, technique, and application. Springer-Verlag, Berlin 27. Sedrak M, Alaminos-Bouza AL, Srivastava S (2020, June 12) Coordinate systems for navigating stereotactic space: how not to get lost. Cureus 12(6):e8578

Influence of the Reorientation Function ...

277

28. Talairach J, Szikla G (1967) Atlas of stereotaxic anatomy of the telencephalon. AnatomoRadiological Studies. Paris: Masson and Cie, p 326 29. Talairach J, Tournoux P (1988) Co-planar stereotaxic atlas of the human brain. 3-Dimensional Proportional System: An Approach to Cerebral Imaging. Stuttgart: Georg Thieme Verlag, p 122 30. Woolrich M, Ripley B, Brady J, Smith S (2000) Nonparametric estimation of temporal autocorrelation in FMRI. Sixth International Conference on Functional Mapping of the Human Brain, p 610

Comparison and Evaluation of Models for Predicting Immunogenicity of Viral Antigens of the pMHC Complex from Murine Models Gracjan Ka˛tek, Marta Gackowska, Karol Harwtig, and Anna Marciniak

Abstract Due to the growing importance of immunotherapy, especially in the treatment of cancer or designing personalized vaccines, there is a need to understand the mechanisms of the adaptive immune response. Based on the available data on the immunogenic response of T cells (CD8 +) to viral peptides presented on the molecules of the major histocompatibility complex (MHC) class I, models of predicting immunogenicity were developed using methods such as: Decision Tree, Support Vector Machine and Extreme Gradient Boosting. Models were compared and validated to choose the best method of predicting immunogenicity. Keywords Prediction model · Machine learning · Immunogenicity · XGBoost · SVM · Decision tree

1 Introduction In recent years, immunotherapies have become a promising method of oncotherapy for some types of cancer [14, 20], due to insufficient traditional methods. Cancer vaccines are aimed at increasing the activity of cytotoxic T lymphocytes specific for the tumor antigen, as well as creating a long-term immunological memory against cancer [23]. Cellular immune responses play an important role in the immune rejection of vascularized tissue in animals and humans [25]. The strength of peptide binding with MHC molecules is one of the most important determinants of protein immunogenicity [16]. An important element in the development of immune therapies are peptide antigens presented to T lymphocytes by the proteins of the major histocompatibility complex (MHC) [31]. Class I MHC molecules are found on all surfaces of nucleated G. Ka˛tek (B) · M. Gackowska · K. Harwtig University of Science and Technology in Bydgoszcz, Bydgoszcz, Poland e-mail: [email protected] A. Marciniak Nicolaus Copernicus University in Toru´n, Ludwik Rydygier Collegium Medicum in Bydgoszcz, Bydgoszcz, Poland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_28

279

280

G. Ka˛tek et al.

cells and present endogenous antigens to the immune system [18]. The interaction of MHC class I with CD8 + molecules on cytotoxic T cells enables the destruction of infected cells by initiating apoptosis. CD8 + T cells are considered to be the major anti-tumor effectors and have been shown to contribute to tumor regression [2]. Class II MHC interacts with CD4 on the surface of helper T cells and is involved in the formation of acquired immunity. It is the responsibility of the MHC to recognize non-host cells and present them to T lymphocytes, and to recognize host cells to prevent autoimmune reactions. During presentation, the MHC molecule binds to the T cell receptor as well as to CD4 + or CD8 + (depending on the type of T cells). The antigen epitope is retained in the groove of the MHC molecule and interacts with the Ig-Like TCR variable domain. This leads to the activation of T cells. In the case of viruses, studies indicate [8] that for peptides presented by MHC class I proteins, the peptide binding affinity is the best indicator of immunogenicity. It has been shown [5] that T cells are endowed with strong viral peptide recognition, and targeting immunogenic neoantigens to related T cells is the most promising and pace-limiting approach in the development of personalized TCR-T cell immunotherapy [21]. Moreover, T lymphocytes are the main adaptive immune cells that directly induce apoptosis of neoplastic cells [12, 24]. Using IT technologies, prediction models are created, which, among other things, identify antigen-specific epitope sites [7] and T-lymphocyte epitopes [22]. These models make a significant contribution to the prediction of T-cell epitopes in tumors [4]. Machine Learning methods are a promising technology used in medicine. It was shown in [26] that most of the machine learning algorithms can reliably predict peptide/MHC binding affinity. The authors in [27] develop and compare three improved SVM (support vector machine) models for cancer prediction and identification and compare them with XGBoost and SVM. The available clinical data in humans are limited, however, there are animal data, including studies in mice. There is a data set of T-cell (CD8 +) immunogenic responses to viral peptides displayed on murine major histocompatibility complex (MHC) class I molecules as measured by cytotoxicity and IFNγ release assays. Based on the available data showing the amino acid sequences in the epitope, MHC type and treatment outcome, we built and compared peptide-MHC (pMHC) immunogenicity prediction models. Thanks to it, with some probability, we are able to predict whether the immune system will start fighting cancer cells and give the right answer or not. The paper is organized as follows: Sect. 2 describes the materials and methods used to develop immunogenic response prediction models, Sect. 3 discusses and compares the effectiveness of the prediction methods developed, and Sect. 4 summarizes the results presented in this paper.

2 Material and Method The data used in the study have been obtained from The Immune Epitope Database (IEDB) [28]. We used over 7,000 records, each of which consisted of 3 main columns: epitope, MHC, and immune response. The epitope columns consist of

Comparison and Evaluation of Models for Predicting Immunogenicity ...

281

peptide sequences from 8 to 11 amino acids in length. MHC columns have 6 types of MHC (H2-Db, H2-Kb, H2-Kk, H2-Kd, H2-Dd, H2-Ld) and the immune response is 1 or 0. Our experiment consisted of the following steps: (i) data preparation and processing, (ii) development of prediction models using machine learning, (iii) training and validation of the developed models.

2.1 Data Preparation and Processing In the first stage, data was prepared and processed in such a way as to properly start teaching and validating models. Label encoding was used to transform the area of interest (MHC) into unique label encoding. As a result of using label encoding, the labels are converted to numeric form, and each amino acid is assigned a unique number.

2.2 Machine Learning Prediction Model Based on the available and prepared data, three prediction models were developed using machine learning methods: Decision Tree, SVM (Support Vector Machine) and XGBoost (Extreme Gradient Boosting). The methods were implemented in Python 3 using the scikit learn library, along with the preprocessing and metric tools. In addition, the xgboost library, math and numpy were used. From the sklearn library were used the svm tools and decision trees to implement the tested methods. The Support Vector Machine (SVM) [19] is machine learning model capable of performing linear, non-linear, regression and outlier detection [11]. SVM is especially useful in the problem of complex data classification, where we have a small or medium set of data. At the core of the algorithm is the concept of decision space, which is divided by building objects belonging to different classes. Decision tree is a popular machine learning algorithm where the analogy to trees occurring in the natural environment is noticeable. They consist of three main elements of the trunk, that is the tree attribute, the branches which are decisions and the leaves which represent the classes [10, 13]. The Fig. 1 presents a diagram of the tree used by the authors of the article. Due to its high complexity, it has been limited to level 3. This method can be used in both regression and classification problems. In the case of this research, the problem of classification has been addressed. Due to the way decision trees work, there are many methods of their implementation. The most important of them are presented in Table 1. The decision tree in this paper was generated using the implementation CART (classification and regression tree) algorithm. The CART algorithm was used in this work, as it is one of the most popular and most frequently used methods of building a decision tree. This approach builds a binary tree by splitting the records at each node

282

G. Ka˛tek et al.

Fig. 1 Scheme of the part of Decision Tree Table 1 Methods of decision tree implementation Algorithm name Data type Classifier CART

Categorical, Numerical

ID3 (Iterative Dichotomiser 3)

Categorical

C4.5

Categorical, Numerical

Description

Gini Index

Numerical division is used to build the tree [13] Entropy and The method needs Information Gain attributes in a discrete form, it cannot cope with continuous attributes [1] Improved ID 3 version It can work on both discrete and continuous data [9]

according to the function of a single input field. Moreover, due to the mechanism of decision trees operation, the criteria of division such as: Gini index, Entropy and Information Gain were used. The Gini index is defined by a formula: Gini = 1 −

n  ( pi )2

(1)

i=1

where pi is the probability of classifying the object in the specified class, and n is the total number of i classes.

Comparison and Evaluation of Models for Predicting Immunogenicity ...

283

XGBoost (Extreme Gradient Boosting) [6] is another machine learning algorithm based on the decision tree, which uses a gradient increase structure. It is one of the latest machine learning methods for supervised learning problems, where we use training data (with many functions) to predict the target variable.

2.3 Training and Validation of Prediction Model The dataset has been divided into two parts: train dataset and test dataset. A set of 15% of the data was assigned to the tests, while the remaining part was used as training data. In each validation trial, the datasets consisted of a different number of positive and negative samples. A 10-fold and a 5-fold cross-validation was performed to calculate the performance indicators. In the study, the confusion matrix [29] presented in the Table 2 was used as a measure of evaluation. It consists of four elements arranged as follows: TP (true positive), FP (false positive), TN (true negative) and FN (false negative). Base on the confusion matrix, the performance index were expressed as follows equation: TP (2) Pr ecision = T P + FP TP T P + FN

(3)

Pr ecision × Recall Pr ecision + Recall

(4)

Recall = F1 = 2 × Fβ = (1 + β 2 ) × G − Mean =

β2 √

Pr ecision × Recall × Pr ecision + Recall

(5)

Pr ecision + Recall

(6)

AU C(Ar ea U nder T he R OC Cur ve) =

1 × (Pr ecision + Recall) 2

MCC(Matthew corr elation coe f f icient) T P × T N − FP × FN =√ (T P + F P)(T P + F N )(T N + F P)(T N + F N ) Table 2 Confusion matrix Actual positive Actual negativ

Predicted positive

Predicted negative

TP TN

FP FN

(7)

(8)

284

G. Ka˛tek et al.

3 Result The purpose of this section is to evaluate the performance and effectiveness of machine learning algorithms in a 5-fold and 10-fold validation process. Several measures of performance were used for the results obtained. All models were learned and validated on the same data sets. The results presented in the Table 3 correspond to the average values obtained in a 5-fold cross-validation for the three machine learning algorithms described in the study. The Fig. 2 shows the results of the prediction algorithms based on the F1 measure, precision and accuracy obtained during 5-time cross-validation. XGBoost achieves the best results, which shows its highest performance compared to SVM and Decision Tree. The Table 4 shows the results of the prediction algorithms on the murine MHC protein immunogenicity dataset with respect to a 10-fold cross-validation. The presented results are the average values of the indicators obtained in 10-fold cross-validation. As seen, the XGBoost algorithm shows the highest efficiency among the three models. The Fig. 3 shows the results of the algorithms on the basis of the F1 measure as well as precision and accuracy obtained during 10-fold cross-validation. SVM and Decision Tree produce similar results, while XGBoost produces the best results. This allows for its greatest effectiveness to be stated, as was the case with the results obtained in the 5-fold validation process.

Table 3 Results of 5-fold cross-validation Decision tree Pr ecision weighted Standard deviation F1weighted Standard deviation Accuracy Standard deviation F1micr o Standard deviation F1macr o Standard deviation

0.7186 0.0348 0.7108 0.0350 0.6971 0.0427 0.7042 0.0340 0.5477 0.0496

SVM

XGBoost

0.6416 0.0043 0.7125 0.0037 0.8010 0.0027 0.8010 0.0027 0.4447 0.0008

0.7842 0.0377 0.7770 0.0300 0.8157 0.0217 0.8157 0.0217 0.6005 0.0584

Comparison and Evaluation of Models for Predicting Immunogenicity ...

285

Fig. 2 The result of prediction 5-time cross-validation for Decision Tree, SVM, XGBoost Table 4 Results of 10-fold cross-validation Decision tree Pr ecision weighted Standard deviation F1weighted Standard deviation Accuracy Standard deviation F1micr o Standard deviation F1macr o Standard deviation

0.7141 0.0263 0.7091 0.0261 0.7036 0.0382 0.7085 0.0375 0.5586 0.0475

SVM

XGBoost

0.6416 0.0043 0.7125 0.0037 0.8010 0.0027 0.8010 0.0027 0.4447 0.0008

0.7842 0.0377 0.7770 0.0300 0.8157 0.0217 0.8157 0.0217 0.6005 0.0584

Both for the 5-fold and 10-fold cross-validation, the immunogenicity prediction model developed with the XGBoost algorithm is characterized by the best parameters compared to the other methods. It is worth noting that Decision Tree has the worst results for each of the cross-validations. The Fig. 4 shows the error matrices for the prediction algorithms developed. For the SVM, 1466 positive samples and 373 negative samples were correctly predicted, the remaining samples were determined incorrectly. For XGBoost, 1,523 positive samples and 316 negative samples were correctly predicted. For Decision Tree, 1321 are positive samples and 518 are negative samples. The ROC curve was also used to evaluate the effectiveness of the proposed methods. The quality of the classification is determined by the two indices AUC and the Gini index. In our case, by evaluating the proposed methods, reading the data from

286

G. Ka˛tek et al.

Fig. 3 The result of prediction 10-time cross-validation for Decision Tree, SVM, XGBoost

(a) SVM

(b) Decision Tree

(c) XGBoost Fig. 4 Confusion matrix

Comparison and Evaluation of Models for Predicting Immunogenicity ...

287

the curve presented in the Fig. 5, we can define the most effective XGBoost algorithm. In turn, the decision trees in this case tend to AUC = 0.5, which allows to call this method a random classifier. This is also confirmed by the values of the gini coefficient presented in the Table 5, which for the decision tree algorithm reach the value of GC = 0.16. The Table 6 shows the results for three indicators: G-Mean, MCC and Fβ . The value of the β index in the proposed solution was set at the value of β = 2. These results show that, as in the previous sections, where the authors discussed the 5-fold and 10-fold cross-validation results, XGBoost achieved the best results. Comparing the memory consumption in the methods proposed by the authors, we can see that the xgboost method consumes the least memory. The decision tree algorithm is in second place and svm needs the most memory. SVM methods and decision trees have similar values in terms of the time needed to perform the operation, xgboost differs significantly from the others. However, taking into account the results

Fig. 5 Receiver Operating Characteristic (ROC) curve

Table 5 The result of gini coefficient for implemented methods Method Gini coefficient (GC = 2 × AU C − 1) SVM Decision tree XGBoost

0.32 0.16 0.52

Table 6 The result of G-Mean, MCC and Fβ for implemented methods G-Mean Fβ MCC

Decision tree

SVM

XGBoost

1.2041 0.7286 0.1693

1.1969 0.6624 0.0

1.2805 0.7938 0.3565

288

G. Ka˛tek et al.

Table 7 Comparison of time and memory consumption in the implemented methods Decision tree SVM XGBoost Time consumption [s] 5.5752 Memory consumption 143.46 [Mb]

4.9916 215.04

15.0656 133.12

obtained and the use of both time and memory, it can still be concluded that xgboost is the best method (Table 7). After the analysis of the state of art and most frequently used methods in [17] and the newer [15] for the described methods of prediction of T-cell epitopes and prediction of HLA peptide binding, the XGBoost method used in this paper was not indicated. It is a new and promising prediction method showing the best results among the methods for ∼82% accuracy and AUC = 0.76. Additionally, in [3] the SVM reaches accuracy of ∼80%, and in our model it also has the value of ∼80%. On the other hand, the decision tree is now a rarely used algorithm. The model developed by the authors reaches accuracy ∼70%, while in [30] the model achieves accuracy of 80%. This may be due to the smaller number of test and validation samples.

4 Conclusion Machine learning models, as presented in the paper, are a promising tool to assess the probability of a cellular immune response. This could lead to the development of an appropriate immunotherapy for cancer. The research presented by us proves that the best effectiveness, efficiency and precision is achieved by the prediction model developed on the basis of XGBoost. In turn, the worst prediction model is based on the Decision Tree. Thus, the immune response prediction model based on this method is characterized by the lowest accuracy and the lowest validity. While the immunogenicity prediction model was designed and validated in murine models, future research will focus on tumor neoantigens based on human clinical data.

References 1. Beel J, Langer S, Genzmehr M, Nürnberger A (2013) Introducing Docear’s research paper recommender system. In: Proceedings of the 13th ACM/IEEE-CS joint conference on Digital libraries - JCDL 2013. ACM Press, New York (2013) 2. Boon T, Coulie PG, Van den Eynde B (1997) Tumor antigens recognized by T cells. Immunol Today 18(6):267–268 3. Bhasin M, Raghava GPS (2004) SVM based method for predicting HLA-DRB1*0401 binding peptides in an antigen sequence. Bioinformatics 20(3):421–423

Comparison and Evaluation of Models for Predicting Immunogenicity ...

289

4. Brusic V, Bajic VB, Petrovsky N (2004) Computational methods for prediction of T-cell epitopes-a framework for modelling, testing, and applications. Methods 34(4):436–443 5. Calis JJA, Maybeno M, Greenbaum JA, Weiskopf D, De Silva AD, Sette A, et al (2013) Properties of MHC class I presented peptides that enhance immunogenicity. PLoS Comput Biol 9(10):e1003266 6. Chen T, Guestrin C (2016) XGBoost: a scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining - KDD 2016. ACM Press, New York 7. Cohen T, Moise L, Ardito M, Martin W, De Groot AS (2010) A method for individualizing the prediction of immunogenicity of protein vaccines and biologic therapeutics: individualized T cell epitope measure (iTEM). J Biomed Biotechnol 2010:1–7 8. Croft NP, Smith SA, Pickering J, Sidney J, Peters B, Faridi P et al (2019) Most viral peptides displayed by class I MHC on infected cells are immunogenic. Proc Natl Acad Sci USA 116(8):3112–3117 9. Freund Y, Mason L (1999) The alternating decision tree learning algorithm. Inicml 99:124–133 10. Gershman A, Meisels A, Lüke KH, Rokach L, Schclar A, Sturm A (2010) A decision tree based recommender system. InIICS 170–179 11. Géron A (2017) Hands-on Machine Learning With Scikit-Learn and TensorFlow: Concepts Tools and Techniques to Build Intelligent Systems. O’Reilly, Sebastopol 12. Hanson HL, Donermeyer DL, Ikeda H, White JM, Shankaran V, Old LJ et al (2000) Eradication of established tumors by CD8+ T cell adoptive immunotherapy. Immunity 13(2):265–276 13. Jadhav SD, Channe HP (2016) Efficient recommendation system using decision tree classifier and collaborative filtering. Int Res J Eng Technol 3:2113–2118 14. Jiang S, Good D, Wei MQ (2019) Vaccinations for colorectal cancer: Progress, strategies, and novel adjuvants. Int J Mol Sci 20(14):3403 15. Kar P, Ruiz-Perez L, Arooj M, Mancera RL (2018) Current methods for the prediction of T-cell epitopes. Peptide Sci 110(2):24046 16. Lazarski CA, Chaves FA, Jenks SA, Wu S, Richards KA, Weaver JM et al (2005) The kinetic stability of MHC class II: peptide complexes is a key parameter that dictates immunodominance. Immunity 23(1):29–40 17. Luo H, Ye H, Ng HW, Shi L, Tong W, Mendrick DL et al (2015) Machine learning methods for predicting HLA-peptide binding activity. Bioinf Biol Insights 9(Suppl 3):21–29 18. Morera Y, Sánchez J, Bequet-Romero M, Selman-Housein K-H, de la Torre A, HernándezBernal F et al (2017) Specific humoral and cellular immune responses in cancer patients undergoing chronic immunization with a VEGF-based therapeutic vaccine. Vaccine 35(28):3582– 3590 19. Noble WS (2006) What is a support vector machine? Nat Biotechnol 24(12):1565–1567 20. Ribas A, Butterfield LH, Glaspy JA, Economou JS (2003) Current developments in cancer vaccines and cellular immunotherapy. J Clin Oncol 21(12):2415–2432 21. Riley TP, Keller GLJ, Smith AR, Davancaze LM, Arbuiso AG, Devlin JR, et al (2019) Structure based prediction of neoantigen immunogenicity. Front Immunol 10:2047 22. Roggen EL (2008) Models for prediction of immunogenicity. In: Immunogenicity of biopharmaceuticals. Springer, New York, pp 75–95 23. Rosenberg SA, Yang JC, Restifo NP (2014) Cancer immunotherapy: moving beyond current vaccines. Nat Med 10(9):909–915 24. Rosenberg SA (2001) Progress in human tumour immunology and immunotherapy. Nature 411(6835):380–384 25. Schriber H (2003) Tumor immunology. In: Paul WE (ed) Fundamental immunology. Lippincott Williams & Wilkins, Philadelphia, pp 1557–1592 26. Smith CC, Chai S, Washington AR, Lee SJ, Landoni E, Field K et al (2019) Machine-learning prediction of tumor antigen immunogenicity in the selection of therapeutic epitopes. Cancer Immunol Res 7(10):1591–1604 27. Turki T, Wei Z (2018) Boosting support vector machines for cancer discrimination tasks. Comput Biol Med 101:236–249

290

G. Ka˛tek et al.

28. Vita R, Mahajan S, Overton JA, Dhanda SK, Martini S, Cantrell JR et al (2019) The immune epitope database (IEDB): 2018 update. Nucleic Acids Res 47(D1):D339–D343 29. Zaki MJ, Meira W Jr (2020) Data mining and machine learning: Fundamental concepts and algorithms, 2nd edn. Cambridge University Press, Cambridge 30. Zhu S, Udaka K, Sidney J, Sette A, Aoki-Kinoshita KF, Mamitsuka H (2006) Improving MHC binding peptide prediction by incorporating binding data of auxiliary MHC molecules. Bioinformatics 22(13):1648–1655 31. Zinkernagel RM, Doherty PC (1974) Restriction of in vitro T cell-mediated cytotoxicity in lymphocytic choriomeningitis within a syngeneic or semiallogeneic system. Nature 248(5450):701–702

Finite Element Analysis of Lumbar Disc Implant, in Aspect of Treating Degenerative Changes in Spine Dawid Ke˛szycki, Bogdan Dybała, Grzegorz Ziółkowski, and Patrycja Szymczyk-Ziółkowska

Abstract Degenerative changes in lumbar spine are one of the most common reason for lower back pain occurring. Unfortunately there are no any known methotds to reverse this process. That is why there are many approaches to find an optimal solution for treating these conditions. In the article numerical analysis of designed implant is presented as a tool for design evaluation, and its optimization according to results of a numerical experiment. Obtained numerical results allowed to point areas with stress concentration and changing their geometry to avoid this. Keywords Finite element analysis · Intervertebral disc implants

1 Introduction Nowadays, more and more people are suffering from pain symptoms in the lower back [14]. One of the causes is degeneration process of the intervertebral disc. Discs in the human spine are structures witch provide proper amortization of loads, and make whole structure flexible, and mobile. Degenerative processes causes, that height of intervertebral disc is decreasing, which can results in vertebras collapsing. It not only may reduce movement range, but also press the nerves and cause sudden, acute pain. That is why current implantology tries to develop implants, which will reset discs functions and reduce pain symptoms. With each passing year, custom implants are more and more common, because of their great fitting to specific patients [1]. Developing this kind of implants is long process, which includes many various aspects, like fitting implants geometry into patient, optimization, analysis, manufacturing, etc. [9, 15]. Additive manufacturing allowed to design implant, which will fit to the specific patient in best possible way. Currently there are many works, where authors presents their idea of designing implant [3, 4]. This article is focused on verifying and D. Ke˛szycki (B) · B. Dybała · G. Ziółkowski · P. Szymczyk-Ziółkowska Politechnika Wrocławska, Wydział Mechaniczny, Katedra Technologii Laserowych, Automatyzacji i Organizacji Produkcji, Wrocław, Poland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_29

291

292

D. Ke˛szycki et al.

optimization process of proposed implant for lumbar disc. Verification of the project was prepared by using finite element analysis (FEA) which is powerful tool for initial evaluation because it only requires 3D models, and knowledge about boundary conditions like, material, fixation and load definition.

2 Designing Custom Made Implant Designing implant requires using reverse engineering methods. First step in preparing such implant is getting imaging data from the patient in DIACOM format, most often CT or MRI images. Thanks to proper software for editing that data, it is possible to rebuild model of patients tissues. Having that model allows to design implant in way that fits modeled structures in the best way. In Fig. 1 there are presented three concepts of intervertebral disc implant. The first concept was prepared as two-elements implant, with ellipsoidal, convex part in bottom boundary plate, which is fitting into concave part of upper boundary plate. That geometry provide moving range in all axes, which also occurs in physiological spine. Both upper plates are textured to provide better contact between bone and implant. Plates also have spikes, which are supposed to stick into bone tissue, and thanks to that implant is better fixed in the bone. One of the drawbacks of this solution is fact, that this type of connection is not providing load dissipation. The second concept is three-elements implant, with two the same boundary plates with intermediary element, which supposed to be made from plastic. That type of implant provides better load dissipation than implant from first concept. It has got similar spikes and texture of boundary plate to provide better fixation of the implant. The third concept is also consisting three elements. Although intermediary element in this concept has different shape than in second concept. It is sliding on bottom plate and can provide rotation between upper plate. The idea to prepare lumbar intervertebral disc implant came from amount of patients whose suffers from LBP, possibility to use AM techniques and a chance to make parametric implant, which can be easily rebuild to fit patients. At first imaging data from 3 different patients was acquired and proceeded to obtain 3D model of each patient’s spine in L3–L5 segment. Each of the model was measured to receive characteristic dimensions for each patient, which were used later, in designing implant.

(a) First concept Fig. 1 Implant concepts

(b) Second concept

(c) Third concept

Finite Element Analysis of Lumbar Disc Implant ...

293

1) Load depreciation - better solution is this one, which provide dissipating incoming loads. This could be obtained by using some elastic materials. Better load dissipation is lowering the risk of damaging spine structures. 2) Amount of elements - when implant is more compact and having less number of elements it is harder to make mistakes during surgery. 3) Complication of surgery - the less additional features the better. If there is no need to use specific tools during surgery or fixating screws, etc., the concept is better. 4) Possibility to fit implant to specific patient - in this criterium easiness of changing implant’s geometry to specific patient was considered better. 5) Possibility to adjust bending angle - geometry which can provide different range of motions in different planes, (as in the physiological spine) is better. Also adjusting movement range to individual patient is beneficial. 6) Possibility to recreate all of the natural movements of the intervertebral column - the more similar movements to natural motion range of spine the better. Each of the concept was rated on a scale from 1 to 5, where 5 was maximum amount of points from single criterium and 1 was minimum. 5 mean that concept fulfil given criterium. After rating concept amount of points was multiplied by weight of criterium. Next all received points were summed up, which gave the final result for each one of the concept. According to presented criteria in Tables 1 and 2 concept II was pointed as the best one. That was the reason to develop implant from this concept.

Table 1 Concepts rating according to prepared criteria Criteria Weight Concept I Load depreciation Amount of elements Complication of surgery Fitting implant to specific patient Possibility to adjust bending angle Recreating all of the natural movements

Concept II

Concept III

3 1

1 5

3 3

2 3

2

3

3

4

4

2

4

3

2

3

4

3

4

3

5

4

294

D. Ke˛szycki et al.

Table 2 Final point result for each concept Concept I Points

40

Concept II

Concept III

62

51

Table 3 Material properties used in FEA Young Modulus [GPa] Titanium alloy [7, 12] UHMWPE [5]

110 7.5

Poisson’s ratio [–] 0.33 0.46

2.1 Boundary Conditions Material, which was proposed for considered implant was titanium alloy Ti6Al4V this alloy is commonly used in additive manufacturing methods, which were considered to be target manufacturing processes for this implant. Moreover, this material is commonly used in implantology due to its good biological behavior, like biocompatibility, and low elastic modulus, which can prevent from stress-shielding appearance. Intermediary element was proposed to be made from ultrahigh molecular weight polyethylene (UHMWPE). This kind of polymer is strong enough to withstand load in lumbar spine and at the same time its elastic enough to provide amortization. For numerical experiment’s simplification all materials were considered as isotropic. Mechanical properties like Poisson’s ratio and Young’s modulus were presented below in Table 3. Printed titanium’s stiffness depends on filling so it could be decreased by creating openwork construction. For the experiment purposes value of totally filled Ti6Al4V alloy young modulus was used. Numerical analysis requires load and fixation definition. For the purpose of presented research load was considered as weight of body segments above the lumbar spine, force was applied to upper surface of L4, vertically, causing vertebras compression. Fixation was applied to bottom surface of L5 vertebra. These conditions also were simplified, cause in the physiological spine, there are more forces, for example those caused by abdominal pressure etc. Although this case was evaluated as the most critical, and proper for implant’s geometry gauging. Load value was calculated as mass of body segments above lumbar spine and equal 450 N. This value was calculated using Zaciorski’s factors for 90 kg male. Load and fixation definition was presented in the Fig. 2.

Finite Element Analysis of Lumbar Disc Implant ...

295

Fig. 2 Boundary conditions used in FEA

Fig. 3 Deformation results in whole implant

2.2 Results of Numerical Experiment Numerical analysis was prepared in ANSYS v18.2. Finite element used in the experiment was 10-node tetrahedral element. This kind of the element is functioning well in irregular geometries like bones. In every picutre there is nodal sollution presented. All stress values were given in MPa - because of the complexitiy of the boundary conditions material tension state was described using Huber-Mises theory. Displacement results were given in mm. Above were presented deformation results for implant’s numerical analysis. In Fig. 3 was presented results for whole implant. Maximum values can be observed for upper boundary plate, where they reach value of 0,064 mm. Maximum values in this part of implant were caused by boundary conditions, and its values suggested that implant was properly fixated and material will not deform uncontrollably.

296

D. Ke˛szycki et al.

As it comes to stress values for both boundary plates this values oscillated between 70 MPa (as presented in Fig. 4). That value is much lower than tensile strength for used titanium alloy manufactured using additive manufacturing which equals between 0,9–1,45 GPa [11]. That indicated that risk of damaging implant in this case is insignificant. Although the fact that stress values were not causing risk of failure it was decided to change geometry of boundary plate to avoid stress concentration in sharp areas on the surface, to minimalize risk of fatigue. As shown in Fig. 5 maximum stress values in absorbing element made from UHMWPE. Maximum value equals 17,4 MPa. That value is lower than strength of the material which is about 40 MPa. That indicates that this element is not endangered by damage.

Fig. 4 Von-Mises stress in implant

Fig. 5 Von-Mises stress in UHMWPE element

Finite Element Analysis of Lumbar Disc Implant ...

297

Fig. 6 Optimized shape of the implant

2.3 Optimization The main purpose of optimization process was proposing geometry of implant, which will be strong enough to survive in human spine, resisting huge forces, which are caused not only by weight of the human body, but also some overloads caused by daily activities. That is why, it was decided to change geometry of boundary plates even though stress level, gained in numerical experiment was not dangerous for implant functionality. New geometry design was presented in Fig. 6. The goal of changing geometry was to avoid sharp finishing of boundary plates surfaces. In new design, it has got circular pockets, which provides better stress distribution. Changing geometry not only decreased maximum stress to 57 MPa, but also as it can be seen in Fig. 6 it also provided more homogenous stress distribution, which is beneficial in aspects of implant’s durability and its contact quality with bone tissue.

3 Conclusion Known methods for treatment intervertebral disc diseases are oriented to relieve pain from patients. There are a lot of works, in which authors are trying to answer the question, which way of treating this type of condition is the best for the patient [1, 2, 10]. If compared to physiological spine functions, movable implants seems to be the most proper prosthesis. There are still carried out researches which will provide longterm results of functioning that kind of implants. It is necessary to clearly answer the question if that kind of treatment is the best. Contemporary medicine is also trying to find the best way to approach disabled segment, which also has a lot of influence

298

D. Ke˛szycki et al.

in the surgery success [13]. Moreover surgery itself should be prepared properly like including imagination devices, which will reduce possibility of making mistakes, and what comes with it increase implant durability [6]. Even small displacement between implant and vertebra body can lead to implant’s failure. That fact makes individual implants better cause of lower probability of movement. Despite of the fact, that TDR process seems to provide results similar to conditions prevailing in the human organism, there are some doubts if such complicated procedure is necessary to provide good movability to patients. There are many works, which include researches in this subject, but there is still no unequivocal answer which of treatment’s method is the best for the intervertebral disc degeneration process [16]. For sure there are still needed results in long-term behavior of movable implants. Currently available results indicate that best results for TDR is achieved for young patients. At the same time pain level measured for patients with movable and stiff implants is similar [14]. Presented finite element analysis provided enough data to evaluate proposed implant design. Final approval of concept should not only require more specific numerical analysis, which will consider not only one static case, but also dynamic problems and effects from other tissues. There are however authors, who prepared FEA for the implant [8]. Their results lead to similar conclusion that FEA alone resulted that implant’s geometry is proper and it will not fail under the mechanical conditions. Moreover FEA needs to be validated by actual experiments made in laboratory. That is why, if project would be developed further it would be necessary to prepare actual experiments to validate FEA. Developing this concept, based on numerical analysis still needs to be considered in holistic way. There is not only one proper way of designing good working implant, so all possibilities should be verified, like choosing other materials, changing implant’s geometry etc. It might cause, that the costs of project will increase significantly, but it could turn out to be crucial concept for current medicine.

References 1. de Beer N, van der Merve A (2013) Patient-specific intervertebral disc implants using rapid manufacturing technology. Rapid Prototyp J 19:126–139. https://doi.org/10.1108/ 13552541311302987 2. Calvo-Echenique A, Cegoñino J, Chueca R, Pérez-del Palomar A (2018) Stand-alone lumbar cage subsidence: a biomechanical sensitivity study of cage design and placement. Comput Methods Prog Biomed 162:211–219. https://doi.org/10.1016/j.cmpb.2018.05.022 3. Doicin C, Ulmeanu ME, Frîncu AS, Enache VC (2017) Concept development of a new lumbar intervertebral disk implant. In: MATEC web of conferences, vol 137, p 02002. https://doi.org/ 10.1051/matecconf/201713702002 4. Domanski J, Skalski K, Grygoruk R, Mróz A (2015) Rapid prototyping in the intervertebral implant design process. Rapid Prototyp J 21:735–746. https://doi.org/10.1108/RPJ-09-20130096 5. Fang L, Leng Y, Gao P (2006) Processing and mechanical properties of ha/uhmwpe nanocomposites. Biomaterials 27:3701–3707. https://doi.org/10.1016/j.biomaterials.2006.02.023

Finite Element Analysis of Lumbar Disc Implant ...

299

6. Janssen M, Garcia R, Miller L, Reed W, Zigler J, Ferko N, Hollmann S (2017) Challenges and solutions for lumbar total disc replacement implantation. Spine 42:S108–S111. https://doi.org/ 10.1097/BRS.0000000000002454 7. Karolewska K, Ligaj B, Wirwicki M, Szala G (2019) Strength analysis of ti6al4v titanium alloy produced by the use of additive manufacturing method under static load conditions. J Mater Res Technol 9:1365–1379. https://doi.org/10.1016/j.jmrt.2019.11.063 8. Karpi´nski R, Jaworski R, Szala L, Ma´nko M (2017) Influence of patient position and implant material on the stress distribution in an artificial intervertebral disc of the lumbar vertebrae. In: ITM Web of conferences, vol 15, p 07006. https://doi.org/10.1051/itmconf/20171507006 9. Mattei T, Beer J, Teles A, Rehman A, Aldag J, Dinh D (2017) Clinical outcomes of total disc replacement versus anterior lumbar interbody fusion for surgical treatment of lumbar degenerative disc disease. Global Spine J 7:452–459. https://doi.org/10.1177/2192568217712714 10. Mobbs RJ, Coughlan M, Thompson R, Sutterlin CE, Phan K (2017) The utility of 3D printing for surgical planning and patient-specific implant design for complex spinal pathologies: case report. J Neurosurg Spine SPI 26(4):513–518. https://doi.org/10.3171/2016.9.SPINE16371 11. Murr L, Quinones S, Gaytan S, Lopez M, Rodela A, Martinez E, Hernandez D, Martinez E, Medina F, Wicker R (2009) Microstructure and mechanical behavior of ti-6al-4v produced by rapid-layer manufacturing, for biomedical applications. J Mech Behav Biomed Mater 2:20–32. https://doi.org/10.1016/j.jmbbm.2008.05.004 12. Niinomi M (1998) Mechanical properties of biomedical titanium alloys. Mater Sci Eng A 243:231–236. https://doi.org/10.1016/S0921-5093(97)00806-X 13. Pokorny G, Marchi L, Amaral R, Jensen R, Pimenta L (2019) Total disc replacement by the lateral approach - up to 10 years follow-up. World Neurosurg 122:e325–e333. https://doi.org/ 10.1016/j.wneu.2018.10.033 14. Salzmann S, Plais N, Shue J, Girardi F (2017) Lumbar disc replacement surgery - successes and obstacles to widespread adoption. Curr Rev Musculoskel Med 10:153–159. https://doi. org/10.1007/s12178-017-9397-4 15. Siu T, Rogers J, Lin K, Thompson R, Owbridge M (2018) Custom-made titanium 3-dimensional printed interbody cages for treatment of osteoporotic fracturee-related spinal deformity. World Neurosurg 111:1–5. https://doi.org/10.1016/j.wneu.2017.11.160 16. Stubig T, Ahmed M, Ghasemi A, Nasto L, Grevitt M (2017) Disc replacement versus anteriorposterior interbody fusion in the lumbar spine and lumbosacral junction: a cost analysis. Global Spine J 8:129–136. https://doi.org/10.1177/2192568217713009

Controlling of the Upper Limb Prosthesis Using Camera and Artificial Neural Networks Agata Mrozek, Martyna Sopa, Jakub K. Grabski, and Tomasz Walczak

Abstract The loss of the upper limb, especially the hand, can affect the level of autonomy. Developing an effective control system for the upper limb prostheses could improve the quality of users’ life. The aim of this project was to design artificial neural networks for automatic grasp classification. A subset of the grips allowing to perform everyday activities was proposed. The proposed artificial neural networks were evaluated and the maximal accuracy reached 97%. Keywords Prosthetic hand · Artificial neural networks · Convolutional neural networks · Grasp classification

1 Introdution Due to hand’s complex anatomy and the variety of movement, controlling of the upper limb prosthesis using electromyography (EMG) signals could be ineffective. The conditions of optimal control are fulfilled when numerous EMG signals acquired from different muscle groups are generated both independently and, as necessary, simultaneously. The EMG signals used to control the prosthesis are measured on the skin surface. This results in crosstalk phenomenon. The electrodes’ positions have strong influence on the quality of recorded signals. The EMG signal is stochastic and vulnerable to disturbances. This causes errors in grasp selection [3, 4]. The important factors influencing the grasp kinematics are object being manipulated, its size, as well as its weight, and task to be performed. The same or similar way to hold certain type of object is used. It allows to predict the applied grasp based on its shape data. People can recognize the characteristics features of object based only on its two-dimensional representation [5, 6, 12–14]. A. Mrozek · M. Sopa · J. K. Grabski (B) · T. Walczak Institute of Applied Mechanics, Poznan University of Technlogy, Poznan, Poland e-mail: [email protected] A. Mrozek e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_30

301

302

A. Mrozek et al.

Convolutional neural networks (CNNs) found special application in image classification tasks. They can extract features from object’s geometry, which can determine the grasp type to be applied. In study descibed in [9], a system based on deep learning was developed. The 2-layers CNN’s architecture was applied. In the first tests, the classification accuracy reached 85% for objects, which were included in the training set, and 75% for objects, which weren’t in the data used for network training. The classification accuracy in the test on real objects reached 84%. Then, two patients after upper limb amputation tested created system. The obtained classification accuracy was equal to 88% [9]. In another example of using camera in order to automatic grasp classification, CNN named VGG-VeryDeep-16 was applied. It enables 1000 object to be identified. Only last layer of network was modified to take into account five grasp types. Based on images captured from camera placed on the prosthesis surface, the achieved grasp classification accuracy was 93.2% [4]. The aim of this study was to design artificial neural networks for automatic grasp classification.

2 Methods 2.1 Applied Grasp Subset During designing prosthetic appliances, it is important to find a compromise between the size of the device and the restored functionality. Therefore, the appropriate subset of grasp to be implemented in the prosthesis should be indicated [11, 15]. Based on the results of the research [1, 2, 7, 16, 17], a subset of grips allowing to perform everyday activities was proposed. This comprises 4 types of grips (the names of the grips in accordance with to the nomenclature of GRASP taxonomy): medium wrap, precision disc, palmar pinch and tripod.

2.2 Image Databse The applied database includes images of everyday objects. It was created on the basis of two image databases: Amsterdam Library of Objects Images (ALOI) and Newcastle Grasp Library [8, 9]. Based on [10], selected images from ALOI were divided into four groups: medium wrap, precision disc, palmar pinch and tripod. The ALOI and Newcastle Grasp Library were combined in order to generate the database, which consists of 39 836 photos.

Controlling of the Upper Limb Prosthesis Using Camera and Artificial Neural Networks

303

2.3 Image Preprocessing Image preprocessing could enhance obtained results. During the study, the impact of 3 commonly used preprocessing techniques, i.e., mean normalization, standardization, zero component analysis (ZCA) was examined. The results of applying different image preprocessing techniques are shown in Fig. 1. The data partition were performed by using hold-out techniques. It partitioned the original dataset into a training set (70%) and a testing set (30%).

3 Results The first step of creating new CNN is to define a network architecture, which depends on the number and types of layers used. Based on the analysis of the existing networks’ structure, 19 sample architectures were created. The general scheme of them

Fig. 1 Image preprocessing a mean normalization, b standardization, c zero component analysis

304

A. Mrozek et al.

is shown in Fig. 2. The input image size is 48 × 36 pixels. It was decided to use the max pooling layer, which preserves the most relevant elements in the image. The selected activation function is the Rectified Linear Unit (ReLU). The number of applied convolutional layers ranged from 1 to 4. Tested architectures were characterised by different filters size and numbers of filters, see Tab. 1. In convolutional layers, option ‘padding’ was also taken into account. The size of added padding was calculated by the software at training. In our analysis, we defined and calculated: – sensitivity Sensitivit y =

TP , T P + FN

(1)

Speci f icit y =

TN , T N + FP

(2)

– specificity

– accuracy Accuracy =

TP +TN , T P + T N + FP + FN

(3)

– precision - positive prediction value (PPV) P PV =

TP , T P + FP

(4)

TN , T N + FN

(5)

– negative prediction value (NPV) N PV =

where T P, F P, T N and F N are numbers of true positives, false positives, true negatives and false negatives, respectively. The obtained classification accuracy was in the range of 77.30 to 93.93% (see Fig. 3). It was noticed that the accuracy above 90% was obtained by networks, which had 3 or 4 convolutional layers. In addition, a significant increase in classification accuracy with the increase in the number of network parameters was noticeable. Two networks, which achieved the best classification accuracy, were chosen for further simulations. Selected networks were tested with different filter sizes and using various preprocessing techniques: mean normalization, standardization, ZCA method. The obtained classification accuracy is shown in Tables 2 and 3. An increase in classification accuracy with the increase of filters size was observed. The highest value of classification accuracy was obtained for network architecture which filters size was equal to 7 × 7 pixels. In the further analysis, standarization technique was applied as the method of preprocessing, because it resulted in the best obtained accuracy.

Controlling of the Upper Limb Prosthesis Using Camera and Artificial Neural Networks Table 1 Summary of tested architectures Network number

Convolutional layer number

Filter size

1

1

5×5

5

–

2

1

5×5

5

–

2

5×5

5

–

3

1

3×3

8

‘same’

4

1

3×3

8

‘same’

2

3×3

16

‘same’

1

3×3

8

‘same’

2

3×3

16

‘same’

3

3×3

32

‘same’

1

3×3

8

‘same’

2

3×3

16

‘same’

3

3×3

32

‘same’

4

3×3

64

‘same’

7

1

5×5

5

‘same’

8

1

5×5

5

‘same’

2

5×5

25

‘same’

1

5×5

5

‘same’

2

5×5

25

‘same’

3

5×5

125

‘same’

1

5×5

5

‘same’

2

5×5

25

‘same’

3

5×5

125

‘same’

4

5×5

625

‘same’

11

1

5×5

20

‘same’

12

1

5×5

20

‘same’

2

5×5

20

‘same’

1

5×5

20

‘same’

2

5×5

20

‘same’

3

5×5

20

‘same’

1

5×5

20

‘same’

2

5×5

20

‘same’

3

5×5

20

‘same’

4

5×5

20

‘same’

1

5×5

5

‘same’

2

5×5

5

‘same’

1

5×5

5

‘same’

2

5×5

5

‘same’

3

5×5

5

‘same’

1

5×5

5

‘same’

2

5×5

5

‘same’

3

5×5

5

‘same’

4

5×5

5

‘same’

18

1

5×5

5

–

2

5×5

25

–

19

1

5×5

5

–

2

5×5

25

–

3

5×5

125

–

5

6

9

10

13

14

15 16

17

Number of applied filters

Padding (in convolutional layer)

305

306

A. Mrozek et al.

Fig. 2 General scheme of the tested convolutional neural networks Table 2 Comparison of classification accuracy for networks with different filter sizes Network number 9 (%) 10 (%) Filter size [3 × 3] Filter size [5 × 5] Filter size [7 × 7]

86.74 92.60 94.06

89.89 93.93 95.51

Controlling of the Upper Limb Prosthesis Using Camera and Artificial Neural Networks

307

Fig. 3 The obtained classification accuracy a for all tested convolutional neural networks, b for tested convolutional neural networks after scaling of the y-axis Table 3 Comparison of classification accuracy for networks with different image preprocessing techniques Network number Filter size Mean Standarization ZCA (%) normalization method (%) (%) 9 9 10 10

5×5 7×7 5×5 7×7

90.63 92.74 94.18 93.60

92.98 94.37 93.82 96.40

93.17 94.50 93.32 94.50

The impact of available optimization methods on the obtained results was checked. MATLAB software has 3 built-in typed of optimization algorithms: ‘sgdm’, ‘rmsprop’, ‘adam’. The additional kinds of layer, like dropout and batch normalization layer, have been taken into account. The use of ‘rmsprop’ and ‘adam’ optimization methods allows to achieve better classification accuracy–especially in the networks, where batchNormalizationLayer and dropout technique were applied, see Table 4.

308

A. Mrozek et al.

Table 4 Comparison of classification accuracy for networks with different optimizer Network number 10 10 10 Filter size Batch normalization layer and dropout layer Optimizer Accuracy [%]

7×7 YES

7×7 YES

7×7 YES

‘sgdm’ 93.35

‘rmsprop’ 96.65

‘adam’ 97.15

Table 5 Determined values of sensitivity, specificity, positive predictive values (PPV) and negative predictive values (NPV) Grasp type Sensitivity (%) Specificity (%) PPV (%) NPV (%) Precision disc Medium wrap Palmar pinch Tripod

96 99 99 95

98 100 100 98

96 100 99 96

98 100 100 98

Based on the performed simulations, the network, which classification accuracy in the test reached 97.15%, was determined. The network no. 10 with filter size equals 7 × 7 pixels and additional layers (batchNormalizationLayer and dropoutLayer) was chosen to implement. The ‘adam’ optimizer was used during learning process. For the selected architecture sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were determined, see Table 5. Selected CNN achieved the best classification results for palmar pinch and medium wrap. Tripod is the grasp type for which the lowest sensitivity value was noted. The most frequently observed error was to classify object for which precision disc should be used into tripod object group.

4 Conclusions The aim of this project was to design an artificial neural network for automatic grasp classification. Proposed grasps subset was created based on the frequency of particular types of grip-manipulation activities occurrence. Individual user’s preferences should also be taken into account. During the determination of the artificial neural network, the impact of the applied image preprocessing technique on the results was compared. It was observed that the network, when standardization of images was carried out, obtained higher accuracy than in the case of the ZCA method and the mean normalization. During the analysis of the obtained results, it was noticed that the number of convolutional layers affects the accuracy of classification. The accuracy above 90% was observed when applying 3 or 4 convolutional layers. Moreover,

Controlling of the Upper Limb Prosthesis Using Camera and Artificial Neural Networks

309

the increase of the number of parameters resulted in the increase of the classification accuracy. The results were also affected by the applied size of filters. Based on the performed simulations, we propose using the CNN in grasp types clasification, because of the accuracy (97% in the best case). The control of myoelectric prostheses could become more effective by using proposed approach based on automatic grasp selection. The application of designed CNN allows to extend the possibilities of offered prosthetic appliances and significantly facilitates the process of switching the type of grasp. Acknowledgements The work was supported by the grant 0612/SBAD/3567 funded by the Ministry of Higher Education and Science, Poland.

References 1. Bullock IM, Feix T, Dollar AM (2013) Finding small, versatile sets of human grasps to span common objects. In: 2013 IEEE International Conference on Robotics Automation, pp 1068– 1075 2. Bullock IM, Zheng JZ, Rosa SDL, Guertler C, Dollar AM (2013) Grasp frequency and usage in daily household and machine shop tasks. IEEE Trans Haptics 6(3):296–308 3. Das N, Nagpal N, Bankura SS (2018) A review on the advancements in the field of upper limb prosthesis. J Med Eng Technol 42:532–545 4. DeGol J, Akhtar A, Manja B, Bretl T (2016) Automatic grasp selection using a camera in a hand prosthesis. In: 2016 38th annual international conference of the IEEE engineering in medicine and biology society, pp 431–434 5. Feix T, Bullock IM, Dollar AM (2014) Analysis of human grasping behavior: correlating tasks. IEEE Trans Haptics Objects Grasps 7(4):430–441 6. Feix T, Bullock IM, Dollar AM (2014) Analysis of human grasping behavior: object characteristics and grasp type. IEEE Trans Haptics 7(3):311–323 7. Feix T, Romero J, Schmiedmayer H, Dollar AM, Kragic D (2016) The GRASP taxonomy of human grasp types. IEEE Trans Hum. Mach. Syst. 46(1):66–77 8. Geusebroek J, Burghouts GJ, Smeulders AWM (2004) The Amsterdam library of object images. Int J Comput Vis 61:103–112 9. Ghazaei G, Alameer A, Degenaar P, Morgan G, Nazarpour K (2017) Deep learning-based artificial vision for grasp classification in myoelectric hands. J Neural Eng 14(3):36025–36043 10. Ghazaei G, Alameer A, Degenaar P, Morgan G, Nazarpour K (2017) Deep learningbased artificial vision for grasp classification in myoelectric hands- supplementary material. J Neural Eng 14:36025–36043. https://iopscience.iop.org/1741-2552/14/3/036025/media/ JNE_Ghazaei_aa6802_Supp_data.pdf, Accessed 12 Jan 2021 11. Jochumsen M, Niazi IK, Dremstrup K, Kamavuako EN (2015) Detecting and classifying three different hand movement types through electroencephalography recordings for neurorehabilitation. Med Biol Eng Comput 54(10):1491–1501 12. Klatzky RL, McCloskey B, Doherty S, Pellegrino J, Smith T (1987) Knowledge about hand shaping and knowledge about objects. J Motor Behav 19(2):187–213 13. Kothari A, Morrow J, Thrasher V, Engle K, Balasubramanian R, Grimm C (2018) Grasping objects big and small: human heuristics relating grasp-type and object size. In: 2018 IEEE international conference on robotics and automation, pp 4237–4242 14. Lau M, Dev K, Dorsey J, Rushmeier H (2016) Learning a human-perceived softness measure of 3D virtual objects. In: ACM Transactions on Applied Perception (SAP 2016), pp 65–68 15. Popovic DB, Sinkjaer T (2008) Central nervous system lesions leading to disability. J Autom Control 18(2):11–23

310

A. Mrozek et al.

16. Vergara M, Sancho-Bru JL, Gracia-Ibáñez V, Pérez-González A (2014) An introductory study of common grasps used by adults during performance of activities of daily living. J Hand Ther 27(3):225–234 17. Zheng JZ, De La Rosa S, Dollar AM (2011) An investigation of grasp type and frequency in daily household and machine shop tasks. In: 2011 IEEE international conference on robotics and automation, pp 4169–4175

Lateral Tibial Condyle Fracture Stabilization—A Numerical Analysis Olimpia Promirska and Jakub Słowinski ´

Abstract In this article, four variants of tibial lateral condyle split fracture stabilization (AO 41-B1.1) with two canulated cancellous screws were studied. The impact of a diameter and length of the thread on the distribution and values of the stress and displacement in the stabilization model was analyzed utilizing the finite element method. Geometric model was derived from CT imaging of 39 years-old female tibial bone. The geometric and numerical models of the individual variants of fracture stabilization were prepared in ANSYS Workbench Software. During the analysis, the Huber-Mises stress of each part of the models and the displacement between bone and split fragment were designated. An influence of the thread diameter on maximum values and distribution of the stress was observed while the thread length did not render any significant impact. For all analyzed variants, the displacement values and distribution were similar.

1 Introduction Tibial plateau fracture, classified as Schatzker type I or AO 41-B1.1, is an interarticular fracture with a lateral condylar split. A vertical load action creates a shearing force, which in combination with bending force leads to the fracture. It affects mostly younger people, since a high density of their bone tissue in this area and lack of diseases prevents bone depression from happening [1, 2]. Main goal of tibial plateau fracture treatment is to obtain the joint stability, congruity of articular surface, early motion, as well as correct axial and rotational limb alignment [2, 3]. Tibial plateau fracture may be also undisplaced or displaced. Because of possible complications, it is recommended to use surgical treatment in order to achieve a successful fixation in the latter case [2]. Gold standard of lateral condyle fracture treatment is a method O. Promirska (B) · J. Słowi´nski Faculty of Mechanical Engineering, Department of Mechanics, Materials and Biomedical Engineering, Wrocław University of Science and Technology, 50-370 Wroclaw, Poland e-mail: [email protected] J. Słowi´nski e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_31

311

312

O. Promirska and J. Słowi´nski

of an open reduction and internal fixation (ORIF), using mostly plates and screws to achieve the stabilization [3]. Alternatively, one could use a different percutaneous or arthroscopically-assisted method (ARIF), achieved mainly by utilizing the leg or cancellous screws, providing similar results [4, 5]. In comparison, the cancellous leg screw fixation shows a lower mechanical stability and interfragmentary compression, but according to the researchers neither locking plate nor screw show any clinically relevant interfragmentary motion therefore cancellous leg screws can be used to stabilize the fracture of a healthy, non-osteoporotic bone [6, 7]. Despite the advantages of the locking plate fixation, the stabilization using only two cancellous screws with corresponding washers is recommended as sufficient for young people [2, 8, 9]. In this study the focus was laid on the stabilizing the lateral plateau fracture using only two cancellous screws and corresponding washers. Some studies about forementioned stabilization method were already performed in the literature, i.e. the comparison between this and other approaches [6, 7], the effect of using full threaded screw [10] and introducing a third screw [11] on fixation strength and stability. It was also reported that the geometric parameters of the thread (length, pitch and minor/major diameter) have a significant impact on the value of the pullout force [12, 13]. Based on this research, it could be stated that the geometric parameters of the thread also influence the stress in the screws and bone or the interfragmentary motion, but to our best knowledge, no studies describing such phenomena has been published. In this article, a finite element method (FEM) analysis of the impact of a diameter and length of the thread on the distribution and values of the stress occurring in the cannulated cancellous screws, the washers, and the bone has been performed. Moreover, the displacement of the split fragment in case of the described stabilization method has been studied and presented.

2 Materials and Methods Geometric model of the tibial bone was based on DICOM files derived from CT imaging (1.25 mm cut). The patient was a 39 years-old female with morphologically correct bone tissue. Tibial bone model consisted of two parts—cortical and cancellous. Area of each type of tissue was extracted with InVesalius software and two three-dimensional models were created. Tibial model included a proximal epiphysis and a bone shaft fragment. Split fracture of lateral condyle was simulated using ANSYS SpaceClaim Direct Modeler software by separating the fragment of the condyle with a plane perpendicular to the long axis of bone. Geometric models of four types of cannulated cancellous screws were created according to manufacturer specifications using ANSYS SpaceClaim Direct Modeler. In case of this fracture the 6.5 mm diameter screws are commonly analyzed in literature [6, 7, 13]. To define the impact of screw thread geometry on stress of the implant and the bone and interfragmentary displacement, it was decided to compare the screws with two different diameters—slightly smaller (5 mm) and larger (7 mm). Two thread lengths were selected for comparison—16 and 32 mm since they are also frequently analyzed by

Lateral Tibial Condyle Fracture Stabilization ...

313

Table 1 Dimensions of the analyzed cannulated cancellous screws Screw length Thread length Thread Pitch (mm) (mm) (mm) diameter (mm) Variant 1 Variant 2 Variant 3 Variant 4

60

16 32 16 32

(a) Fracture stabilization model

5 5 7 7

1.77

Core diameter (mm) 3 3 4.8 4.8

(b) Canulated cancellous screw

Fig. 1 Fracture stabilization model analyzed in this article (Variant 1): method of load application and constraint (a), geometric model of cannulated cancellous screws (b)

researchers in this type of fracture [6, 11, 13]. Remaining dimensions of analyzed screws are presented in Table 1. Fracture stabilization was achieved by placing two identical cancellous screws and matching washers as shown on Fig. 1. Each screw was leaning upward and forward respectively by 0.63◦ and 13.62◦ for anterior one and 1.98◦ and 4.16◦ for posterior. Distance between the heads of each screw was 21.77 mm horizontally and 0.97 mm vertically. This arrangement was obtained as a result of an additional analysis aimed at the best adjustment of the washers to the lateral surface of bone. Prepared models were imported to ANSYS Mechanical Enterprise software and discretized. To mesh all the parts, the tetra- and hexahedral elements were used. The number of nodes in the prepared models ranged from 494125 (Variant 1) to 660587 (Variant 4), while the number of elements—in between 305277 (Variant 1) and 383305 (Variant 4). Material models of the screws, the washers and the bone—both cancellous and cortical—were homogeneous and isotropic. Material parameters—Young modulus and Poisson ratio assumed for cortical bone were equal to 18000 MPa and 0.3, respectively; cancellous bone—450 MPa, 0.42, respectively; Ti6Al4V screws and washers—105000 MPa,

314

O. Promirska and J. Słowi´nski

0.34, respectively [14]. Interactions between individual parts of model were defined as a frictional contact, with the exception of the interaction between the washers and the cortical part of split fragment and between cortical and cancellous bone. In the first case, the contact was defined as “no separation” and in the second—“bonded”. The numerical thread was used instead of the geometrical one in order to increase the efficiency of the numerical analysis [15]. Since it is not recommended after the surgery to allow the weight bearing for 10–12 weeks [2, 8], it was decided to use the load value corresponding to that occurring during movement with two orthopedic crutches for 87 kg person. The load of 105.5 N in X axis, −2.3 N in Y axis and −2.4 N in Z axis was applied to each condyle as shown in Fig. 1 [16] imitating standing on both legs [17]. Additionally, in the unthreaded part of each screw the force of 200 N was applied to simulate the holding power of the screws. Distal part of the tibial shaft was constrained without displacement in all axes as shown in Fig. 1. The numerical analysis was performed using ANSYS Workbench software. During the analysis, the Huber-Mises stress of each element of the fracture fixation models and the displacement between bone and split fragment were designated.

3 Results 3.1 Stress Distribution According to the results of finite element analysis, a stress distribution was similar for all models of screw fixation system, however slight differences related to thread diameter can be observed as shown on Figs. 2 and 3. The highest stress values observed for the 5 mm diameter model were concentrated on the washer’s surface (near the central hole) interacting with the lateral surface of the split fragment. On the 7 mm model, however, the highest stress values were localized on the washer’s surface interacting with the head of the screw. Maximum value was observed in the model of stabilization by screws with 5 and 16 mm thread diameter and length, respectively and it was equal to 333.03 MPa. The highest values for each part of all models are presented in Table 2. Concentrations of high stress values occurring on the screw models are located near the connections of unthreaded part with the head of the screw and they decrease gradually to the end of screw. For all variants, higher values occurred on the anterior screw. The highest stress value was observed for screws with 5 mm and 32 mm thread diameter length, respectively and it was equal to 74.34 MPa. Distribution of the stress on the screws was shown on Fig. 2. Maximum value of stress for the bone tissue occurred on the lateral surface of the cortical bone and it was equal to 121.39 MPa. It was observed for a model of stabilization utilizing the screws with 5 mm and 16 mm thread diameter and length, respectively. For each model, high stress values concentrated in the area of interaction between the split fragment and the washers as shown on the left of Fig. 3.

Lateral Tibial Condyle Fracture Stabilization ...

315

Table 2 Calculated values of Maximum Huber-Mises stress for individual parts of stabilization model variants and Maximum relative displacement in fracture plane Maximum Huber-Mises stress (MPa) Variant 1 Variant 2 Variant 3 Variant 4 Washers Screws Split fragment Bone Maximum relative displacement (µm)

333.03 65.67 121.39 7.28 24

(a) Variant 1

311.24 74.34 115.67 5.66 25

244.67 36.63 103.19 5.50 23

244.34 37.08 109.09 4.57 24

(b) Variant 3

Fig. 2 Stress distribution for washer and screw models: stabilization variant 1 with 5 mm thread diameter, 16 mm thread length (a) and variant 3 with 7 mm thread diameter, 16 mm thread length (b). Values are presented in MPa

For the cancellous bone tissue high stress values occurred in the areas corresponding to those observed in cortical bone—on the lateral surface of the split fragment. High values were also located inside the holes and around them on the fracture surface. Especially large area of high stress concentrations was observed under the posterior hole as presented on in the center of Fig. 3. The highest value of stress for each model occurred in the area of interaction between the cancellous bone and the screw thread (right side of Fig. 3). Maximum stress was equal to 7.28 MPa and occurred for the stabilization variant with 5 and 16 mm thread diameter and length, respectively.

316

O. Promirska and J. Słowi´nski

3.2 Displacement Distribution Analysis of the displacement of the split fragment in relation to the bone was in the vertical axis, parallel to the fracture surface. The distribution of displacement values was similar for each stabilization model as it is shown on Fig. 4. The value of the relative displacement was lower than 25 µm. It should be noted that all split fragments descended relatively to the bone. High values of the relative displacement occurred in anterior and bottom area of the fracture surface, whereas low values—in the upper, posterior area. Maximum relative displacement in all models was observed below the posterior hole and they were presented in Table 2.

(a) Variant 1

(b) Variant 3 Fig. 3 Stress distribution throughout various parts of the stabilization model: variant 1 with 5 mm thread diameter, 16 mm thread length (a) and variant 3 with 7 mm thread diameter, 16 mm thread length (b). Values are presented in MPa

(a) Tibial bone

(b) Split fragment

Fig. 4 Displacement distribution in the fracture plane in vertical axis (Variant 1): tibial bone (a) split fragment (b). Values are presented in µm

Lateral Tibial Condyle Fracture Stabilization ...

317

4 Discussion According to the result of the finite element analysis, the highest stress for each model of stabilization occurs on the washers. Slightly lower values were observed in the area of interaction between the washers and the cortical bone. The tibia layer of the cortical bone is particularly thin on the lateral side of the lateral condyle. Obtained values of stress exceeded the compressive strength of the cancellous bone which is 0.6–7.8 MPa according to the area of the proximal tibia [18]. Utilizing the screws without the washers could cause a collapse of the screw into the bone tissue. This observation explains the need of applying the washers and providing the best possible position on the bone surface [19]. The finite element analysis showed that for all models of stabilization, the value of relative displacement did not exceeded 25 µm. According to the literature, the intra-articular displacement of 2 mm is a threshold to treat the fractures operatively, therefore it can be stated that stabilization models were prepared correctly [3]. The main goal of this study is to analyze the impact of the diameter and the length of the thread on the distribution and the value of the stress occurring in the models of the lateral condyle split fracture stabilization. Distribution of the stress, was similar for all stabilization models, but it showed a characteristic similarity for variants of stabilization using the screws with the same diameter of the thread. According to the finite element analysis, the influence of thread length on stress value or distribution was not observed. As mentioned earlier, there are some studies in literature which show the influence of thread geometrical properties such as diameter, length and pitch, on the value of the pullout force [12, 13]. It is therefore possible that analyzing different values of length and diameter or using more specific loading cases will result in noticing some more accurate dependencies. There are some limitations related to, inter alia, the bone geometry and material properties. Prepared model did not included the part of the tibial shaft and distal tibia epiphysis. This simplification has been introduced because no stress concentrations were observed in the distal part of diaphyseal. It was also decided that fibula will be omitted in this model and it will be loaded directly, not by distal epiphysis of femur. Another limitation is the application of isotropic and homogeneous material model for the bone tissue. Using the CT images, however, allows for the extraction of the area of the cortical and the cancellous bone and application of appropriate, averaged mechanical parameters for both kinds of tissue. This actions were driven by the simplification of the finite element model, limiting the number of elements and nodes, and thus reducing the system requirements and speeding up the numerical analysis. The main goal of this study is to compare the variants of screw fixation, therefore it was crucial to provide equal conditions for each variant. To our best knowledge, none of the applied simplifications should affect the final result.

318

O. Promirska and J. Słowi´nski

5 Conclusion In this work, the finite element analysis was used to compare four variants of screws differing in length and diameter of the thread used within the fixation system during the stabilization of the lateral condyle fracture. An influence of the thread diameter on maximum values and distribution of the stress for bone, split fragment and implants has been observed. The thread length did not render any significant impact on stress value or distribution, but it is possible that further investigation of different values and loading cases could make other dependencies noticeable. Acknowledgements The calculations were made using the resources of the Wroclaw Centre for Networking and Supercomputing (http://www.wcss.pl), calculation Grant no. 397.

References 1. Duwelius PJ, Templeman DC (1996) The knee: tibial plateau fracture reduction techniques utilizing cannulated screw fixation. Springer, New York, pp 170–188 2. Schatzker J (1987) Fractures of the tibial plateau. Springer, Heidelberg, pp 279–295 3. Prat-Fabregat S, Camacho-Carrasco P (2016) EFORT open reviews, vol 1 4. Baron ML, Cermolacce M, Flecher X, Guillotin C, Bauer T, Ehlinger M (2019) Orthop Traumatol Surg Res 105:2 ´ T, Ceroveˇcki T, Curi´ ´ c S, Vidovi´c D (2017) Injury 48:11 5. Elabjer E, Benˇci´c I, Cuti 6. Kojima K, Gueorguiev B, Seva G, Stoffel K, de Oliveira RG, Eberli U, Nicolino T, Lenz M (2015) Medicine 94:1 7. Carrera I, Gelber PE, Chary G, González-Ballester MA, Monllau JC, Noailly J (2016) Int Orthop 40:10 8. Hansen M, Pesantez R: Treatment of partial articular fracture, split 9. Rudran B, Little C, Wiik A, Logishetty K (2020) Br J Hosp Med 81:10 10. Salduz A, Birisik F, Polat G, Bekler B, Bozdag E, Kilicoglu O (2016) J Orthop Surg Res 11:12 11. Moran E, Zderic I, Klos K, Simons P, Triana M, Richards RG, Gueorguiev B, Lenz M (2017) J Orthop Transl 11:10 12. DeCoster TA, Heetderks DB, Downey DJ, Ferries JS, Jones W (1990) J Orthop Trauma 4:6 13. Chapman JR, Harrington RM, Lee KM, Anderson PA, Tencer AF, Kowalski D (1996) J Biomech Eng 118:8 14. Słowi´nski JJ (2010) Analiza stanu napre˛˙ze´n w konstrukcji indywidualnego implantu kostnego. Ph.D. thesis, Politechnika Wrocławska, Instytut Materiałoznawstwa i Mechaniki Technicznej. (in Polish) 15. Inzana JA, Varga P, Windolf M (2016) J Biomech 49:6 16. Orthoload: Database 17. Be˛dzi´nski R (1997) Biomechanika in˙zynierska, Zagadnienia wybrane. Oficyna Wydawnicza Politechniki Wrocławskiej. (in Polish) 18. Hvid I, Christensen P, Søndergaard J, Christensen PB, Larsen GC (1983) Acta Orthop Scand 54:1 19. Hansen M, Pesantez R: Lag screw technique

Extreme Compression of the Electrocardiographic Signals Using Matching Persuit ´ Sandra Smigiel

Abstract The area of telemedicine, including wireless communication, has an increasingly significant impact on health care. In recent years, an increase in the importance of compression methods has been noticed in many areas of medicine. Especially in the field of storing, processing and remotely transmitting large amounts of data. Different compression algorithms, based on different methods have been described in the literature so far. Few of them are currently used in monitoring systems and telemedicine. This paper discusses the use of the Matching Pursuit (MP) compression method for electrocardiographic signal transmission using LoRa (Long Range) technology. LoRa technology is one of the communication standards of the Internet of Things. An electrocardiogram (ECG) is a diagnostic tool that measures and records the heart’s electrical activity in detail. The ECG is an important tool used to diagnose heart abnormalities. The paper describes research of the effectiveness of the Matching Pursuit algorithm in compressing the ECG signal, using the Dictionary Learning. The signals from the PTB Diagnostic ECG Database were used for the tests. The analyzed signal fragments were divided into 1000 ms parts, which were then resampled and normalized. The learning data was used to create a dictionary of atoms using the Dictionary Learning method. Using the Orthogonal Matching Pursuit algorithm for the fragments of test data, indices of non-zero coefficients were obtained. In the next step, using the developed extreme compression algorithm, a byte table that could be transmitted over the network was obtained. In the research, the byte array was decompressed and the signal was reconstructed. Consequently, the transmitted to the original signal similarity was measured for different parameters of the algorithm. Thanks to the development of technology, remote patient monitoring, consulting and medical care can be more flexible and convenient. Keywords Signal compression · Matching pursuit · ECG

´ S. Smigiel (B) Bydgoszcz University of Science and Technology, Bydgoszcz, Poland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_32

319

320

´ S. Smigiel

1 Introduction The contemporary information society is increasingly seeing a crisis in the availability of healthcare. This state of affairs has two causes. On the one hand, it is related to the current demographic trends. On the other hand, it is based on human lifestyle. An aging society, often unhealthy lifestyle, determines the increase in the number of people who develop the first symptoms of heart disease. In a short time, seemingly harmless problems turn into chronic diseases [1]. From recent years the analysis of statistical data, both in Europe and in the world, shows an increase in the incidence of cardiology. According to the World Health Organization, cardiovascular disease is the leading cause of death worldwide. The projected demographic trends are not favorable [5]. Undoubtedly, an important factor in this process is the strong dependence of the influence of the patient’s age on the causes of cardiac disorders. Cardiovascular diseases require more and more supervision and constant control. This, in turn, translates into an increase in health care costs and is increasingly difficult to implement, taking into account the infrastructure of the medical facility. It is estimated that this condition will cause a serious increase in the demand for various types of medical services, in particular diagnostic [1, 3, 4]. Long-term recording of ECG signals (often longer than 24 h) is necessary to obtain an accurate diagnosis and to detect specific diseases. The solution is digital wireless data transmission technologies. Their implementation opens the possibility for new telemedicine applications ensuring continuous monitoring of the patient’s health, and thus improving the quality of his life [7, 9]. From an economic point of view, storing, processing and transmitting the entire ECG waveform is not efficient. Therefore, specific compression methods have been applied to the ECG recordings to minimize the amount of data. Due to the subject of the work, the focus was solely on the aspect of compression using the Matching Pursuit method for the transmission data via the LoRa technology. The author omitted the issues involving the analysis of ECG from the perspective of heart disease, focusing solely on transmission techniques. This area was left for further development in subsequent works.

1.1 Diagnostics via ECG The World Health Organization (WHO) lists cardiovascular disease (CVD) as the leading cause of death worldwide. Given the seriousness of these diseases, many researchers have focused their research on cardiovascular disease and heart health [1]. Increased efforts in this area of research have led to several advances in screening and diagnostic technology for cardiac function. A wide variety of heart abnormalities are examined and evaluated using a simple, risk-free and inexpensive heart test, known as electrocardiography. Electrocardiography provides the doctor information about the correct amount of oxygen delivered to the heart muscle cells, electrical impulses, heart rate, etc.

Extreme Compression of the Electrocardiographic Signals Using Matching Persuit

321

Electrocardiogram (ECG) is a graphical record of a bioelectric signal generated by the human body during the heart cycle. In medical terms, measures and records the heart’s electrical activity in detail. Graphically, it provides useful information that relates to how the heart works with baselines and waves representing changes in the heart’s tone over a period of time. Each ECG heartbeat signal contains three visible waves: P wave, QRS complex and T wave. The detection of these waves in a short time and their assessment by a specialist is not only a source of information about the patient’s health condition, but also a rescue for patients requiring urgent care [6]. In this perspective, ECG records are used for screening tests, diagnose and monitor several heart conditions. ECG monitoring concerns not only patients who have experienced cardiac arrest, diagnosed ventricular tachycardia or cardiac arrhythmia. It may also be used for diagnosis [16]. The development of various biomedical sensors as part of the wireless infrastructure does not put any restrictions on the transmission of recorded data to a medical facility. Especially when there are medical irregularities and the patient’s health is at risk. This will allow them to react early and possibly save the patient’s life [2, 9]. Many different compression methods have been used in recent decades, which are used in many areas, such as image and signal processing. Each of these algorithms has its inherent advantages and disadvantages. In the case of ECG signals, compression has become an important issue in biomedical signal processing research. ECG signals are recorded from patients for both monitoring and diagnostic purposes. Therefore, keeping them in a computerized version has become extremely important. In addition to these, there are many advantages of ECG compression such as transmission speed of real-time ECG, which is also economical and increasingly used for patient telemonitoring [8, 10].

1.2 Wireless ECG Monitoring Remote health monitoring is becoming less and less inaccessible and more often a normal trend in many countries. The growing elderly population and the tendency to unhealthy lifestyles are contributing to the increase in chronic diseases. This, in turn, increases the need for constant clinical supervision and, consequently, an increase in costs for healthcare. The adoption of a solution for a mixed mode, including outpatient treatment and remote treatment, allows not only to reduce costs but also to avoid regular and frequent visits in the hospital, improving the quality of life of the patient and limiting infectious complications [3, 4]. In the last decade, there has been a “leap” of wired communication systems, which has been replaced by low-cost mobile communication. Data from the UNESCO Institute for Statistics show that up to 90% of people in the world are in the range of a cell phone transmitter. On the other hand, the number of mobile phone users corresponds to the number of literate people on Earth. Given the lack of healthcare professionals and their continued migration to developed economies, it seems urgent to introduce mHealth as a cost-effective model for providing services [1, 11].

322

´ S. Smigiel

Wireless technology is used practically everywhere, where there are devices for transmitting sound, image and data. In medicine, it plays an important role due to the ability to send and receive information related to the patient’s health over long distances. Biomedical information is sent via wired or wireless communication to remote or mobile devices for processing and monitoring anywhere and anytime [12]. Signal processing and patient health assessment are performed in real time. Emergency services are automatically notified of the detection of adverse events, and information about the patient’s health condition can be sent at regular intervals or on request [14]. The possibilities of wireless technology are growing at a fantastic pace. There seems to be no limit to what technology can achieve given the infinite resources. Acceptance of the mobile system for monitoring the health of patients is primarily a technological challenge. Real-time mobile health monitoring has many design challenges, such as medical device energy consumption, transmission cost and coverage, privacy, availability, usability, transmission security and more. It also covers aspects related to mobile devices and their technical specifications. And also remote data transmission techniques [16].

1.3 Objective This article reviews the use of Matching Pursuit in the ECG signal analysis. The aim of the study was to assess the possibility of using the Matching Pursuit algorithm in ECG signal compression for the purposes of transmission using LoRa technology. Extreme compression in this sense is aimed at transmitting the electrocardiographic signal over a distance using a small amount of transmitted data. The undertaken research was aimed at answering whether and to what extent the compression of the ECG signal is possible. This should allow to determine the possibility of transmitting long-distance fragments compressed ECG signal, taking into account the usefulness of such a signal after decompression.

2 Description of the Implemented Method 2.1 Signal Compression Using Matching Pursuit Increasingly, in the field of compression methods are used the matching pursuit (MP) and orthogonal matching pursuit (OMP). Matching Pursuit is an algorithm for finding a match. This technique is based on finding the best match of the dictionary function to the signal under study. The elements of this dictionary are called atoms. OMP is one of a variation of MP [13, 15].

Extreme Compression of the Electrocardiographic Signals Using Matching Persuit

323

OMP is an iterative greedy algorithm that selects at each step the dictionary element best correlated with the residual part of the signal till the stopping criterion of iteration is satisfied. OMP algorithm is an important greedy optimal atom searching algorithm. In this article OMP it is used to compress the ECG signal, more precisely to obtain indexes of atoms with non-zero coefficients. This allows you to save such data in the form of a byte array, as presented later in this article.

2.2 Dictionary Created Using Dictionary Learning Technique As part of the work, the Dictionary Learning technique was used, which consists in creating dictionaries aimed at creating a set of atoms that would be best suited to represent a given type of signal. Contrary to the standard dictionary of Gabor functions, to create such a set you need data—understood as signal samples, which in our case were the ECG signals [13, 15]. The task of the DL technique is to find the sparse set of atoms that make up the dictionary, which best represents the training data. It is a technique that is part of Machine Learning. An example of selected elements of the dictionary created with the help of ECG signal fragments is shown in Fig. 1. The signal value in mV is presented on the ordinate axis, the sample numbers are presented on the abscissa axis.

Fig. 1 The sample atoms from a dictionary created using Dictionary Learning

324

´ S. Smigiel

Fig. 2 An example of building a byte array

2.3 Convert to Byte Array As a result of the operation of the OMP algorithm, atomic indexes and their coefficients are obtained. This information is stored as a byte array. The author’s goal is to save this information using as few bytes as possible while maintaining the best possible signal quality representation. Each pair: the index of an atom and its coefficient are written as k-bit and l-bit integers. The index of the atom is written as an unsigned number. To be as efficient as possible, the number of atoms in the dictionary is a multiple of 2. For example, for a dictionary of 512 atoms, 9 bits are used to write the index. In order to increase the efficiency of writing the coefficients, all non-zero coefficients are normalized so that their values are in the range −1.1. For this purpose, the largest absolute value of all the coefficients is determined, and then this value is raised to the nearest integer. This value is called d. The coefficients are divided by this value and then multiplied by 2(l−1) (l—the number of bits to write the coefficient). The coefficients are stored as a signed integer. The value of d is stored in the byte array as a non-negative integer, followed by successive coefficient indices and coefficients. The Fig. 2 shows an example of recording from a signal using a 9-bit recording of indexes and a 10-bit recording of coefficients. Decompress is analogous to compression. The coefficients are converted to floating point values, divided by the value 2(l−1) , then multiplied by d, and set at the appropriate indices. After this operation, the signal is reconstructed.

3 Experimental Results and Discussion Electrocardiograph records from a cardiological database, consistent with global standards, have been used for the testing of the algorithm. Clinical data used for the purpose of this article originates from the Physiobank Database. The testing of the method has been based on data acquired from the PTB Diagnostic ECG. This database is provided for PhysioNet and has different morphologies ECG signal, from healthy volunteers and patients with different heart diseases. The experiment was carried out in accordance with the diagram shown in Fig. 3.

Extreme Compression of the Electrocardiographic Signals Using Matching Persuit

325

Fig. 3 Diagram of the experiment performed

Fig. 4 Sample signal before compression

The figures show the fragments of the ECG signal for the best and worst compression cases. Sample signal before compression reconstruction can be seen on Fig. 4. Same signal after reconstruction can be seen on Fig. 5 (for 10 non-zero coefficients) and Fig. 6 (for 50 non-zero coefficients).

326

´ S. Smigiel

Fig. 5 Sample signal after reconstruction (the worst compression cases)

Fig. 6 Sample signal after reconstruction (the best compression cases)

The analysis contained within this article has been conducted by examining 549 registered records of different patients. Each sample is represented in 1 channels, with the frequency 1000 Hz. In practice resampling 500 Hz is performed. The duration of evaluated waveforms is 1 s [17].

Extreme Compression of the Electrocardiographic Signals Using Matching Persuit

327

Fig. 7 Dependence of MSE on the number of transmitted atom indexes for different and different amount of training data

Fig. 8 The dependence of the number of transmitted bytes on the number of transmitted atom indexes for a different amount of training data

The samples were divided into two bases: training data and test data. The training data was used to create a dictionary of atoms using the Dictionary Learning technique. Test data, appropriately cut, normalized and resampled, were used to conduct research on the impact of specific compression parameters on the transmission quality. With the help of the Orthogonal Matching Persuit algorithm, the indexes of the atoms with the corresponding coefficients were obtained and then decompressed. This way, the byte arrays that would normally be transmitted were obtained as described in previous section.

328

´ S. Smigiel

Fig. 9 The dependence of the MSE on the number of transmitted bytes for a different amount of training data

Fig. 10 The dependence of the MSE on the compression ratio for a different amount of training data

At the next stage, the signal was decompressed and reconstructed, and the quality of the obtained signal was assessed by MSE (mean square error). In this paper, the author considered the possibility of conducting a reliable test, the blind test. However, the study focused on the first stage of the quality assessment of reconstruction using the MSE method.

Extreme Compression of the Electrocardiographic Signals Using Matching Persuit

329

Fig. 11 The dependence of MSE on the number of transmitted atom indexes for different sizes of a dictionary

Fig. 12 The dependence of the number of transmitted bytes on the number of transmitted indexes of atoms for different sizes of a dictionary

The analysis was performed for a different number of transmitted atom indexes, and thus for a different number of transmitted bytes. The impact they had on MSE was examined. The research also examined the influence of the number of training data for learning the dictionary, the size of the created dictionary and the number of bits to be written on the coefficient.

330

´ S. Smigiel

Fig. 13 The dependence of MSE on the number of transmitted bytes for different sizes of a dictionary

Fig. 14 The dependence of MSE on the compression ratio for different sizes of a dictionary

Figure 7 shows the dependence of MSE on the number of transmitted atomic indexes for a different amount of training data. It shows that for each tested case, the greater the number of training data, the smaller the MSE. However, the amount of training data (Fig. 8) does not affect the number of transmitted bytes, which increases linearly with the number of transmitted indexes of atoms. This causes (Fig. 9) that for a given training set size, the dependence of the MSE on the number of transmitted bytes has the same form as the dependence on the number of transmitted atom indexes. The same trend can be seen in the dependence of MSE on the compression ratio (Fig. 10).

Extreme Compression of the Electrocardiographic Signals Using Matching Persuit

331

Fig. 15 The dependence of MSE on the number of transmitted atom indexes for a different number of bits per factor

Fig. 16 The dependence of the number of transmitted bytes on the number of transmitted atom indexes for different bits per factor

Figure 11 shows the dependence of MSE on the number of transmitted atom indexes for various dictionary sizes. As can be seen in Fig. 12, the larger the dictionaries, the more bits should be devoted to writing the indexes of the atoms, which means that the message has a total of more bytes. On the other hand, MSE decreases with the use of larger dictionaries (Fig. 13), while the total number of transmitted bytes increases. On Fig. 14 you can see that the compression ratio effect on MSE is similar for each dictionary size.

332

´ S. Smigiel

Fig. 17 The dependence of MSE on the number of transmitted bytes for bits per factor

Fig. 18 The dependence of MSE on the compression ratio for bits per factor

Figure 15 shows the dependence of the MSE on the number of transmitted atom indexes for a different number of bits per factor. It clearly shows that using too few bits for the factor causes a significant increase in MSE. Figure 16 shows the dependence of the number of transmitted bytes on the number of transmitted indexes of atoms, while Fig. 17 shows the dependence of MSE on the number of transmitted bytes for the number of bits per factor. The above figures clearly show that there is a certain optimal number of bits for a factor, above which further increasing the number of bits, and hence increasing the accuracy of writing the factors, does not improve MSE, but increases the total message length. Figure 18, which shows the dependence of the MSE on the compression ratio, confirms this observation.

Extreme Compression of the Electrocardiographic Signals Using Matching Persuit

333

4 Conclusions The digital revolution and the rapid growth of smartphones, mobile connectivity and social networks have changed the way we live. It affected practically everyone industry and every aspect of our lives. However, its most important influence can be seen in the world of medicine. The wireless monitoring landscape for disease management is extremely complex and is evolving rapidly. Acceptance of the wireless health monitoring model is a simultaneous improvement in the quality of life for both sick and elderly patients, as well as healthy people. Especially in the field of patients suffering from cardiovascular diseases. This article reviews the effectiveness of the Matching Pursuit algorithm in compressing the ECG signal, using the Dictionary Learning. The obtained research results indicate the necessity to optimally select the amount of training data, their increase contributes to the extended duration of learning with the use of Dictionary Learning. This, in turn, determines the duration of the research. Another factor analyzed was the size of the dictionaries. The results show that it is worth using large dictionaries, notwithstanding the cost of using more bits to write the indexes of atoms. There is an optimal number of bits for the factor, 10 for the case under study. Undoubtedly, the ability of consumers to monitor their health without the participation of a physician is a paradigm shift related to the development of digital technology and the use of mobile devices, but also a challenge in terms of responsible interpretation of the results obtained. The coming years will be decisive. Validation and acceptance of such solutions may contribute to placing mobile devices in outpatient patient monitoring.

References 1. Alwan A (1982) Global status report on noncommunicable diseases 2010. Arch Rat Mech Anal 78:315–333 2. Bansal A et al (2015) Remote health monitoring system for detecting cardiac disorders. IET Syst Biol 9(6):309–314 3. Brunetti ND, Scalvini S, Acquistapace F, Parati G, Volterrani M, Fedele F, Molinari G (2015) Telemedicine for cardiovascular disease continuum: a position paper from the Italian Society of Cardiology Working Group on Telecardiology and Informatics. Int J Cardiol 184:452–458 4. Clifford GD, Clifton D (2012) Wireless technology in disease management and medicine. Annu Rev Med 63(1):479–92 5. Dilaveris PE, Gialafos EJ, Sideris SK, Theopistou AM, Andrikopoulos GK, Kyriakidis M, Gialafos JE, Toutouzas PK (1998) Simple electrocardiographic markers for the prediction of paroxysmal idiopathic atrial fibrillation. Am Heart J 135:733–738 6. Elgendi M, Eskofier B, Dokos S, Abbott D (2014) Revisiting QRS detection methodologies for portable, wearable, battery-operated, and wireless ECG systems. PLoS ONE 9:1–18 7. Farabi H, Rezapour A, Jahangiri R, Jafari A, Kemmak AR, Nikjoo S (2020) Economic evaluation of the utilization of telemedicine for patients with cardiovascular disease: a systematic review. Heart Fail Rev 25(6):1063–1075

334

´ S. Smigiel

8. Farooq M, Pesch D (2018) Analyzing LoRa. A use case perspective. In: IEEE 4th world forum on internet of things (WF-IoT), pp 355–360 9. Faust O, Acharya UR, Ma J, Min LC, Tamura T (2012) Compressed sampling for heart rate monitoring. Comput Methods Programs Biomed 108(3):1191–1198 10. Jha CK, Maheshkumar HK (2017) ECG data compression algorithm for tele-monitoring of cardiac patients. Int J Telemed Clin Pract 2(1):31–41 11. Kohno R, Hamaguchi K, Takizawa H, Li K (2008) R&D and standardization of body area network (BAN) for medical healthcare. In: Proceedings - IEEE international conference on ultra-wideband (ICUWB 2008), Hannover, Germany, vol 3, pp 5–8 12. Lymberis A (2005) Progress in R&D on wearable and implantable biomedical sensors for better health care and medicine. In: Proceedings of the 3rd annual international IEEE EMBS special topic conference on microtechnologies in medicine and biology 13. Pati YC, Rezaiifar R, Krishnaprasad PS (1993) Orthogonal matching pursuit: recursive function approximation with applications to wavelet decomposition. In: Proceedings of the twentyseventh asilomar conference on signals, systems and computers, pp 40–44 14. Teng XF, Zhang YT, Poon CCY, Bonato P (2008) Wearable medical systems for P-health. IEEE Ann Rev Biomedi Eng 1:62–74 15. Tropp JA, Gilbert AC (2007) Signal recovery from random measurements via orthogonal matching pursuit. IEEE Trans Inf Theory 53(12):4655–4666 16. WHO Monica Project Principal Investigators (1988) The world health organization Monica project (monitoring trends and determinants in cardiovascular disease): a major international collaboration. J Clin Epidemiol 41(2):105–114 17. PTB Database. http://www.physionet.org/physiobank/database/ptbdb/

Experimental Study the Blood Flows in a Transparent Models of a Blood Vessels with Bifurcation—Preliminary Report Wojciech Wolanski, ´ Marek Ples, Marta Sobkowiak-Pilorz, Grzegorz Gruszka, Michał Burkacki, Sławomir Suchon, ´ and Marek Gzik

Abstract The aim of the work was to develop method and build a laboratory stand that would allow the study of flows, using PIV (Particle Image Velocimetry) method, in a transparent model of a blood vessel with bifurcation. For the purposes of the research, a transparent model of a blood vessel was also made. The model was prepared with the assumed morphometric parameters of the arteries: internal diameters of the inlet and outlet canals, as well as the bifurcation angle. PIV apparatus consists of a camera, laser with an optical arrangement to limit the physical region illuminated, a synchronizer to act as an external trigger for control of the camera and laser, the seeding particles, the fluid under investigation and appropriate software. The analysis of the obtained data allows for the conclusion that the proposed laboratory stand enables the study of the flow in the blood vessel model. Keywords Vessel model · Laboratory stand · Flow studying

1 Introduction Nowadays, an increasing percentage of people have pathological changes in the circulatory system. Epidemiological studies indicate that factors such as stress, hypertension, nicotinism, low physical activity promote the growth of lesions in blood vessels [6, 8]. Thanks to early diagnosis and the prompt initiation of appropriate treatment, the chances of extending the patient’s life increase. There are many studies that describe numerical simulations of blood flow through the middle cerebral artery (MCA) [1, 3, 9, 10, 12, 13]. The results show that the

W. Wola´nski · M. Ples (B) · M. Sobkowiak-Pilorz · G. Gruszka · M. Burkacki · S. Sucho´n · M. Gzik Faculty of Biomedical Engineering, Silesian University of Technology, ul. Roosevelta 40, 41-800 Zabrze, Poland e-mail: [email protected] W. Wola´nski e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_33

335

336

W. Wola´nski et al.

highest values of wall shear stress (WSS) and pressure occur at the bifurcation area where the vessel wall weakens which in turn results in the risk of an aneurysm. In order to verify the numerical simulations carried out in the first stage [11], a model of the middle cerebral artery was developed and then a laboratory stand was built to study the flows using the PIV (Particle Image Velocimetry) method, which is a non-invasive quantitative measurement of the velocity fields. Using PIV method there is possibility to analyse the velocity field information of fluid motion from the images and visualize it. There are many PIV methods, which can be used for measuring 2 dimensional flows in a plane [5]. According to Murray [7], the energy expended on maintaining the circulation is minimal when the radius of the main stem and the division branch meet the equation: r03 = r13 + r23

(1)

where: – r0 —radius of the main stem, – r1 —radius of the first branch, – r2 —radius of the second branch.

2 Laboratory Stand Construction The aim of the work was to develop method and build a laboratory stand that would allow the study of flows in a transparent model of a blood vessel with bifurcation. The assumption was to achieve the modelling of at least some properties of the real circulatory system of vertebrates (including humans), using technically simple solutions. The stand that allows the simulation of the flow through the artificial blood vessel imitating the artery must enable the creation of a characteristic, pulsating flow of the working medium (distilled water) with the following physical properties: temperature—20 ◦ C, density—998.23 kg/m3 , dynamic viscosity—1.004 mPa*s. Therefore, apart from the working medium reservoir (Fig. 2A) and the artificial blood vessel (Fig. 2B, detailed view on Fig. 1), the stand also includes a peristaltic pump forcing pulsating flow (Fig. 2C). The internal diameter of artificial blood vessel was 3 mm. The reservoir is made of glass with a total volume of 65l. For the experiment, 5l of the working medium was used. Commonly available as a food raw material, the seeds of the opium poppy (Papaver somniferum) were used due to their appropriate and homogeneous dimensions, buoyancy close to neutral and lack of negative impact on the environment and possible waste. Used pump makes it possible to adjust the flow rate and its direction. The working medium circulates in flexible hoses (Fig. 2D). Internal diameter of hoses was 8 mm. The movement of the working medium with markers inside the blood vessel illuminated by the circular lamp (Fig. 2E) is monitored by a high-speed camera (Fig. 2F,

Experimental Study the Blood Flows in a Transparent Models ...

337

Fig. 1 Artificial vessel model

(a) Live view

(b) Scheme

Fig. 2 Laboratory stand

Basler acA1440–200 um). Data acquired with pylon Viewer (version 6.1.5.7395) software in the form of a video recording or a series of images was saved in the computer memory for further analysis. The image recording speed was up to 900 fps.

3 Results The first analyzes of video files containing information about the flow in an artificial blood vessel were made with the use of the Tracker software (version 5.1.5), available under an open source license [2]. This software allows both manual and automatic image correlation using markers—preferably with regular shapes. Figure 4 shows the image of the markers in the form of small plastic particles carried by the flow of the working medium through the vessel. Color indicators show the effectiveness of image correlation, which allows to determine such quantities as the displacement of particles, their speed, acceleration, and also to visualize their trajectory. We determine that the speed during the test was up to 1.6 m/s.

338

W. Wola´nski et al.

Fig. 3 Particle trajectory

Fig. 4 Image of the markers in the form of small plastic particles carried by the flow of the working medium through the vessel captured in Tracker software

During the analysis, there were cases where, due to the pulsating flow, a given particle initially entered one of the bifurcation branches (Fig. 5A, B) only stopped for a short moment (Fig. 5C), and then moved back and entered the other bifurcation branch (Fig. 5D, E). Figure 3 shows this kind of particle trajectory. The described behavior is the result of the pulsating flow of the medium that normally flows towards the bifurcation through one branch and out through two. During each pumping cycle, there is a moment when the system pressure drops, resulting in a short-term reversal of the flow direction. The fluid travels a short distance in the opposite direction, but this is enough for its small volume to reach the other bifurcation branch.

Experimental Study the Blood Flows in a Transparent Models ...

339

Fig. 5 Specificity of movement of the selected particle

4 Summary The analysis of the obtained data allows for the conclusion that the proposed laboratory stand enables the study of the flow in the blood vessel model. The stand can be developed in order to more accurately reflect the other factors—apart from the speed and nature of the flow and the geometry of the model—affecting the operation of the circulatory system. In the first place, it will be ensuring the stabilization of the working medium temperature, accurate pressure measurement at many points in the circulation, etc.

References 1. Brian DS, Yoganandan N, Stineman MR, Gennarelli TA, Baisden JL, Pintar FA (2007) Mechanics of fresh, refrigerated and frozen arterial tissue. J Surg Res 139:236–242 2. Brown D, Cox AJ (2009) Innovative uses of video analysis. Phys Teach 47:145–150 3. Ferruzzi J, Vorp DA, Humphrey JD (2011) On constitutive descriptors of the biaxial mechanical behavior of human abdominal aorta and aneurysms. J R Soc Interface 8:435–450

340

W. Wola´nski et al.

4. Gzik-Zroska B, Joszko K, Wola´nski W, Gzik M (2016) Development of new testing method of mechanical properties of porcine coronary arteries. Information technologies in medicine. In: Pie˛tka E, Badura P, Kawa J, Wie˛cławek W (eds) 5th international conference, ITIB 2016, ´ ˛ski, Poland, 20–22 June 2016. Proceedings, vol 2. Springer, Cham, pp 289–297. Kamie´n Sla bibliogr. 25 poz 5. Majewski W, Wei R, Kumar V (2020) Developing particle image velocimetry software based on a deep neural network. J Flow Vis Image Process 27(1) 6. Mendis S, Puska P, Norrving B (2011) Global Atlas on cardiovascular disease prevention and control. world health organization in collaboration with the world heart federation and the world stroke organization, pp 3–18. ISBN 978-92-4-156437-3 7. Murray CD (1926) The physiological principle of minimum work. I. The vascular system and the cost of blood volume. Proc Natl Acad Sci USA 12(3):207–214 8. Naghavi M, Wang, H, Lozano R, Davis A, Liang X, Zhou M, et al (GBD 2013 Mortality and Causes of Death Collaborators) (January 2015) Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 385 (9963):117–171 9. Schulze-Bauer CA, Holzapfel G (2003) Determination of constitutive equations for human arteries from clinical data. J Biomech 36:165–169 10. Shojima M, Oshima M, Takagi K, Torii R, Hayakawa M, Katada K, Morita A, Kirino T (2004) Magnitude and role of wall shear stress on cerebral aneurysm computational fluid dynamic study of 20 middle cerebral artery aneurysms. Stroke 35:2500–2505 11. Sobkowiak M, Wola´nski W, Kawlewska E, Gzik M, Joszko K, Zimny M, Kaspera W (2018) Simulation of blood flow in arteries for different flow rates. Modelowanie In˙zynierskie, t 35(66):53–61. (in Polish) 12. Wola´nski W, Gzik-Zroska B, Joszko K, Gzik M, Sołtan D (2017) Numerical analysis of blood flow through artery with elastic wall of vessel. In: Gzik M, Tkacz E, Paszenda Z, Pie˛tka E (eds) Innovations in biomedical engineering, proceedings IiBE 2016. Springer, Cham, pp 193–200 13. Wola´nski W, Gzik-Zroska B, Joszko K, Kawlewska E, Sobkowiak M, Gzik M, Kaspera W (2018) Impact of vessel mechanical properties on hemodynamic parameters of blood flow. In: Gzik M, Tkacz E, Paszenda Z, Pie˛tka E (eds) Innovations in biomedical engineering, proceedings IiBE 2017. Springer, Cham, pp 271–278

ARM-200 - Upper Limb Rehabilitation Robot Andrzej Michnik, Mariusz Sobiech, Jakub Wołoszyn, Mirella Urzeniczok, Aleksander Sobotnicki, Rafał Kowolik, and Krzysztof Cygon´

Abstract The disability of the upper limbs limits the performance of daily activities and work, rehabilitation allows in many cases to reduce limitations and in some cases to restore full efficiency. This article presents an ARM-200 upper limb rehabilitation robot, its kinematic solutions and detailed description of components and roles performed in the robot control system. The article also discusses the possible application of the designed system in a rehabilitation process as well as presents differences between the ARM-200 robot and the ARM-100 robot, the work on which was completed in 2009. A. Michnik (B) · M. Sobiech Łukasiewicz Research Network, Institute of Medical Technology and Equipment, ul. Roosevelta 118, 41-800 Zabrze, Poland e-mail: [email protected] M. Sobiech e-mail: [email protected] Faculty of Biomedical Engineering, PhD School, Silesian University of Technology, ul. Akademicka 2a, 44 - 100 Gliwice, Poland J. Wołoszyn · M. Urzeniczok · A. Sobotnicki Łukasiewicz Research Network, Institute of Medical Technology and Equipment, ul. Roosevelta 118, 41-800 Zabrze, Poland e-mail: [email protected] M. Urzeniczok e-mail: [email protected] A. Sobotnicki e-mail: [email protected] R. Kowolik · K. Cygo´n Przedsie˛biorstwo Handlowo-Usługowe Technomex Sp. z o.o., ul. Szparagowa 15, 44-141 Gliwice, Poland e-mail: [email protected] K. Cygo´n e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_34

341

342

A. Michnik et al.

Keywords Robot · Rehabilitation · Kinematic · Control system · Serious games

1 Introduction We are living in an ageing society suffering from civilisation diseases resulting from limited physical activity, stress and injuries sustained in traffic accidents. According to the WHO, the third place on the list of civilisation diseases leading to disability is occupied by brain stroke [1]. An indispensable element in the treatment of brain strokes is rehabilitation, the effectiveness of which strongly depends on the time of its commencement following the occurrence of the stroke [2]. According to the Supreme Audit Office (Najwy˙zsza Izba Kontroli), Poland, only nearly 30% of patients after stroke undergo rehabilitation [3]. Rehabilitation after brain stroke is a laborious process lasting many months. Sometimes, post-stroke rehabilitation may last until the end of patient’s life. The above-presented factors combined with a limited number of physiotherapists lead to a situation where many patients never fully recover and become independent. One of the ways out of the situation could be the use of robots helping physiotherapists in monotonous and exhausting physical work and enabling one physiotherapist to simultaneously provide many patients with physiotherapeutic rehabilitation. An additional advantage is the fact that the robot can also be used as a diagnostic tool as its sensors can measure, e.g. ranges of mobility in given joints or force exerted by individual muscles. The present development and growing availability of increasingly complicated technical solutions as well as the ongoing pandemic of COVID-19 lead to the conclusion that the near future will see the popularisation of robot-based applications. The article presents a robot for the rehabilitation of upper limbs. The primary assumption concerning the design of the ARM-200 upper limb rehabilitation was the obtainment of the most versatile solution enabling the performance of multi-planar movements in the widest possible ranges by the left or right upper limb. Another important assumption was to design the mechanical structure of the robot in a manner making it possible to combine ergonomics, safe use and aesthetics. The predecessor of the ARM-200 rehabilitation robot was a prototypical ARM-100 robot developed at the Institute of Medical Technology and Equipment (currently belonging to the Łukasiewicz Research Network) in the years 2007–2009 [4]. Both robots are presented in Fig. 1. The prototype of the ARM-100 robot enabled the development of primary technical solutions, yet its design was not optimised for aesthetics, the drives of the robot significantly stuck out of its primary body, whereas electronic packages controlling the operation of the drives and groups of wires were without any casing. However, the primary limitation of the ARM-100 robot control system was the lack of possibility of switching between the left and right limb. The primary assumption is that the ARM-200 rehabilitation robot should enable both active and passive rehabilitation training. Active training is performed using a movement tracking mode based on force sensors. By additionally setting the threshold of insensitiveness of the force measurement line, it is possible to adjust

ARM-200 - Upper Limb Rehabilitation Robot

343

Fig. 1 View of ARM-200 and former ARM-100 robot

force which should be applied to a given robot grip so that the robot could enable the movement of the limb. During passive training, the robot can make movement in selected nodes on the basis of angular values which can be numerically defined by the therapist or based on the data contained in a related file. Such data were recorded during the monitoring of the limb located on the robot arm. The robot control system enables an independent setting of operation modes and parameters for each of the nodes. As a result, it is possible to use the robot in complicated training scenarios.

2 Mechanical Design of the Robot Design works involved analyses of various kinematic diagrams of the robot based on previously performed tests [5]. The first model of the robot was based on the kinematic diagram of the ARM-100 robot. However, it was characterised by a limited movement of arm abduction by the drive at the second kinematic pair of the robot arm. If the angle of abduction exceeded 90°, the robot arm collided with the patient’s head. The second concept of the robot was developed in order to increase the range of adduction and that of abduction of the shoulder joint. Within the aforesaid concept, the first drive of the exoskeleton was placed in a position where its axis of rotation was located vertically above the patient’s shoulder. However, because of a problematic movement or, to be precise, the impossibility of moving from the extreme position of external rotation to the internal rotation of the shoulder joint (very important in terms of rehabilitation), the above-presented concept was rejected. Both concepts of robot models are presented in Fig. 2. Finally, the decision was made to return to the

344

A. Michnik et al.

Fig. 2 Visualisation of the first and second model of the ARM-200 robot

kinematic diagram of the first model. This fact involved the modification of the robot structure consisting in the relocation of drive no. 2 outside the area of collision with the patient’s head. In its final version, the ARM-200 robot is composed of the base and the exoskeletal arm, having seven primary and three auxiliary degrees of freedom. The numbers of kinematic pairs with marked degrees of freedom are presented in Fig. 3. The primary degrees of freedom constitute active kinematic pairs enabling movements performed by both the left and right upper limb. Auxiliary degrees of freedom are used to adjust the exoskeleton to a given patient (through the adjustment of the arm height in relation to the ground as well as the length of the arm and forearm segments). The presented kinematic diagram of the ARM-200 robot enables the performance of movements within the following degrees of freedom: 1. 2. 3. 4. 5. 6. 7.

Flexion - hyperextension in the shoulder joint; Abduction - adduction in the in the shoulder joint; Internal - external rotation in the shoulder joint; Extension - flexion in the elbow joint; Supination - pronation in the elbow joint; Palmar - dorsal flexion in the wrist joint; Radial - ulnar flexion in the wrist joint.

ARM-200 - Upper Limb Rehabilitation Robot

345

Fig. 3 Visualisation of the kinematic pairs of the ARM-200 robot with marked degrees of freedom (1–7 primary degrees of freedom and 1’–3’ auxiliary degrees of freedom)

3 Robot Control System The elements of the control system of the ARM-200 robot were designed to perform dedicated tasks performed by the rehabilitation robot. The general control concept was based on the ARM-100 robot control system, yet, because of the presently available element base and significantly more favourable possibilities of presently available components (e.g. microcontrollers), the entire electronics of the robot control system has been designed anew. In the general concept of the robot, the control system is divided into several components presented in Fig. 4. An integral part of the robot control system is a computer (PC) with the supporting software controlling the process of rehabilitation. The computer is connected to the Master module acting as the interface between the PC and Slave nodes of the control system. Each Slave module performs a preset operation mode on the basis of measurement elements and enables the control of the operation of one BLDC-type motor (i.e. DC motor with electronic commutation). In addition, each of the Slave modules enables interaction with the Expander module, extending the functionality of the node with additional functions, including motor brake control or linear actuator control. In the current design, the module of Expander is only needed in three nodes of the system. Another additional module, with which a given node of the robot can be extended, is the 2-channel EMG module measuring electrical activity of muscles by connected electrodes. The communication of the PC software programme with the robot takes place via Ethernet enabling data transfer of 100 Mbps. In terms of the ARM-100 model, communication was via USB 2.0 in Full Speed mode, significantly limiting

346

A. Michnik et al.

Fig. 4 Schematic diagram of the ARM-200 robot control system

the responsivity of communication with the robot. The communication between the PC application nad Slave modules uses UDP packages sent in cycles through the Master module. Each package sent from the computer application is received by the master module. In next step the Master divides the collected data package into small packages directed to appropriate nodes of the system. After sending each package to subsequent node, the Master module waits for a response of the Slave module. Communication between the Master module and Slave modules takes place via a 2wire (half-duplex) RS-485 bus. An additional role of the Master module (presented in Fig. 5) is the control of the module of safety relays, enabling the stopping of the robot drives at any moment (by means of emergency switches or the Master module) in the event of a failure. The another essential element of the robot are Slave modules, controlling the operation of robot drives on the basis of commands and parameters read out from the sensors connected to a given Slave module. The operation modes of Slave modules include the mode of movement tracing and the mode of movement reproduction. The mode of movement tracing enables the performance of active rehabilitation training. In this type of training, an algorithm traces limb

ARM-200 - Upper Limb Rehabilitation Robot

347

Fig. 5 Schematic diagram of the Master module along with power supply and the module of safety relays

movements using force sensors and aims to obtain the value of force equal to zero. In addition, the movement-tracing algorithm enables the setting of loads through setting the threshold of insensitiveness. If movements are recorded, the threshold of insensitiveness is set at the minimum level ensuring the stopping of a given drive if no force is applied to the sensors. In turn, during active training, by increasing the threshold of insensitiveness, it is possible to set greater resistance which should be overcome by the patient. The second primary operation mode of the robot’s nodes is the reproduction of movement, where the robot makes a movement to a given position and at a given velocity. This mode enables the performance of passive training, where a given movement can be set manually or reproduced on the basis of a movement previously recorded in the tracing mode. After switching the robot on, a read-out of the absolute angle value is copied to the encoder of a relative position of the drive (incremented and decremented on the basis of Hall sensors located inside the BLDC motor interacting with the Slave module). The schematic diagram of the Slave module elements is presented in Fig. 6. The primary role of Hall sensors (embedded in the motor) is the control of the commutation process by MOS-FET power transistors. In their calculations, the PID controllers (responsible for the stabilisation of velocity and position) use information read out from a relative encoder controlled by the Hall sensors of the motor. Another very important element of the rehabilitation robot is a software application enabling control the rehabilitation process

348

A. Michnik et al.

Fig. 6 Schematic diagram of the Slave module of the robot control system

through the selection of a scheme of exercises, recording of rehabilitation results and providing patients with attractive training programmes. In the project presented in the article, the PC application (named VAST.rehab) interacting with the robot was developed by the Brontes Processing company, Gliwice. During rehabilitation training, the patient sits opposite a large monitor displaying training games, where, depending on the type of rehabilitation, the patient performs movements affecting the motion of animated objects. Rehabilitation becomes a form of entertainment, which is of particular importance in the cases of laborious and multi-month long rehabilitation processes. Numerous publications confirm a significant effect of games on patient’s motivation when performing rehabilitation exercises of their upper limbs [6]. In reference articles concerning various rehabilitation robots [7, 8], a frequently raised issue was whether such robots enabled providing more effective rehabilitation than the manual one. Presently available test results reveal that rehabilitation robots help achieve comparable rehabilitation effectiveness as that obtainable using manual rehabilitation.

4 Conclusions The above-presented ideas enabled the making of a new rehabilitation robot and solved many mechanical, electronic and IT problems characteristic of its predecessor.

ARM-200 - Upper Limb Rehabilitation Robot

349

In comparison with the previous robot (ARM-100), the mechanical design and the control system of the new robot (ARM 200) have been significantly modified, making the solution more versatile. An example of improved versatility is the possibility of rehabilitation of the left or right upper limb, alternately. The implementation of the aforesaid possibility required the solution of numerous technical and mechanical problems, including the issues connected with the wiring and the robot control system. Another crucial modification involved the use of drives with harmonic gear in four nodes. The new gear type enabled the reduction of mechanical play in the robot arm joints as well as made it possible to reduce dimensions of robot’s driving elements by eliminating additional bearing and shortening gears themselves. Because of the multitude and complexity of crucial elements of such a system, the Authors fully realise the necessity of further development of many elements of the robot. Acknowledgements This research was financed by the National Centre for Research and Development, Poland, within grant No POIR.01.02.00-00-0014/17 and co-financed by the Ministry of Higher Education and Science of Poland within grant No DWD/3/7/2019 - RJO15/SDW/001.

References 1. Johnson W, Onuma O, Owolabi M, Sachdev S (2016) Stroke: a global response is needed. Bull World Health Organ 94:634–634A. http://dx.doi.org/10.2471/BLT.16.181636 2. Wasti SA, Surya N, Stephan KM, Owolabi M (2021) Healthcare settings for rehabilitation after stroke. In: Platz T (eds) Clinical Pathways in Stroke Rehabilitation. Springer, https://doi.org/ 10.1007/978-3-030-58505-1_14 3. The Supreme Audit Office (Najwy˙zsza Izba Kontroli) (2017) Implementation of the tasks of the National Health Fund in 2017, Evidence Number 156/2018/P/18/055/KZD 4. Michnik A, Brandt J, Szczurek Z, Bachorz M, Paszenda Z, Michnik R, Jurkoj´c J (2010) Control system for a limb rehabilitation robot. Adv Intell Soft Comput 69-Inf Technol Biomed 2:423– 430, Berlin 5. Michnik R, Jurkoj´c J, Rak Z, Me˛˙zyk A, Paszenda Z, Rycerski W, Janota J, Brandt J (2008) Kinematic analysis of complex therapeutic movements of the upper limb. Springer-Verlag, Berlin Heidelberg, Inf Technol Biomed, Adv Soft Comput 47:551–558 6. Koutsiana E, Ladakis I, Fotopoulos D, Chytas A, Kilintzis V, Chouvarda I (2020) Serious gaming technology in upper extremity rehabilitation: scoping review. JMIR Serious Games 8(4):e19071. https://doi.org/10.2196/19071 7. Mehrholz J, Pollock A, Pohl M, Kugler J, Elsner B (2020) Systematic review with network meta-analysis of randomized controlled trials of robotic-assisted arm training for improving activities of daily living and upper limb function after stroke. Mehrholz et al. J Neuroeng Rehabil 8. Rodgers H, Bosomworth H, Krebs H, van Wijck F, Howel D, Wilson N, Finch T, Alvarado N, Ternent L, Fernandez-Garcia C, Aird L, Andole S, Cohen DL, Dawson J, Ford GA, Francis R,Hogg S, Hughes N, Price Ch, L Turner DL, Vale L, Wilkes S, Shaw L (2020, October) Robot-assisted training compared with an enhanced upper limb therapy programme and with usual care for upper limb functional limitation after stroke: the RATULS three-group RCT. Health Technology Assessment 24(54), ISSN 1366-5278

Correction to: Porous Structure and Surface Chemistry of Biomorphous Composite Derived from Carbonized Yucca Covered by Thin Film of Chitosan and Its Application to Remove Hazardous Substances Justyna Majewska, Marta Krzesińska, Temenuzhka Budinova, Anna Filipowska, Nartzislav Petrov, and Boyko Tsyntsarski

Correction to: Chapter 20 in: M. Gzik et al. (eds.): Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_20

In the original version of the book, the following belated corrections are to be incorporated: In chapter “Porous Structure and Surface Chemistry of Biomorphous Composite Derived from Carbonized Yucca Covered by Thin Film of Chitosan and Its Application to Remove Hazardous Substances”, the affiliation “Institute of Organic Chemistry, Bulgarian Academy of Sciences, Acad. G. Bonchev Street, BL. 9, 1113, Sofia, Bulgaria” of authors “Temenuzhka Budinova, Nartzislav Petrov, Boyko Tsyntsarski” are to be changed to “Institute of Organic Chemistry with Centre of Phytochemistry, Bulgarian Academy of Sciences, Acad. G. Bonchev Street, BL. 9, 1113, Sofia, Bulgaria. The correction book has been updated with the changes.

The updated version of this chapter can be found at https://doi.org/10.1007/978-3-030-99112-8_20 © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 M. Gzik et al. (eds.), Innovations in Biomedical Engineering, Lecture Notes in Networks and Systems 409, https://doi.org/10.1007/978-3-030-99112-8_35

C1