A Practical Approach to Stereo EEG 9780826136923, 9780826136930, 2020038905, 0826136923

Citation preview

STEREO EEG Stephan U. Schuele, MD, MPH

With chapters written by leading experts from around the world, the book is divided into 10 sections covering noninvasive evaluation, technical aspects, electrode planning, practical approach for specific epilepsies, surgical placement in adults and children, interpretation, brain mapping, surgical procedures, and outcomes. Chapters integrate highlighted key concepts with illustrative case examples throughout to enhance clinical applicability. Four detailed case discussions of specific epilepsy syndromes covered in the book are also available online to demonstrate the process of patient evaluation, surgical planning, and decisionmaking in a multidisciplinary patient management conference. A Practical Approach to Stereo EEG is the essential comprehensive clinical handbook for practitioners at any level of training or experience involved in invasive EEG evaluations or working at surgical epilepsy centers.



STEREO EEG

Stereo EEG has revolutionized the way invasive EEG explorations are performed, facilitating the assessment of more complex cases with increased precision, a lower surgical risk, and better patient outcomes. A Practical Approach to Stereo EEG is the first dedicated reference on stereoelectroencephalography written for trainees, physicians, and technologists involved in invasive EEG evaluation and monitoring. This go-to resource provides a practical overview of the concepts, methodology, technical requirements, and implantation strategies for common and uncommon surgical epilepsies amenable to stereo EEG. Including over three hundred detailed figures, anatomical drawings, and MRI correlations, this guidebook is an indispensable tool for anyone training, practicing, and teaching in the field.

A Practical Approach to

A Practical Approach to

An Imprint of Springer Publishing

A Practical Approach to

STEREO EEG

Key Features: n

n n n

 orld-class contributors with global expertise provide hands-on experience in successful use W of stereo EEG in complex situations Additional online chapter-based narrated cases discuss specific epilepsy syndromes Purchase includes access to ebook available on most devices and computers

Shelving Category: Neurology

An Imprint of Springer Publishing

11 W. 42nd Street New York, NY 10036-8002 www.springerpub.com

Schuele

n

 overs all practical aspects of stereo EEG, including methodology, technical requirements, and C strategies to successfully perform and interpret invasive monitoring Highly illustrated cases are interwoven within chapters to heighten clinical use

Stephan U. Schuele

A Practical Approach to Stereo EEG

A Practical Approach to Stereo EEG Editor

Stephan U. Schuele, MD, MPH Professor of Neurology, Physical Medicine and Rehabilitation Feinberg School of Medicine Northwestern University; Epilepsy Section Head Department of Neurology Medical Director Neurological Testing Center Northwestern Memorial Hospital Chicago, Illinois

Copyright © 2021 Springer Publishing Company, LLC Demos Medical Publishing is an imprint of Springer Publishing Company. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Springer Publishing Company, LLC, or authorization through payment of the appropriate fees to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400, fax 978-646-8600, [email protected] or on the Web at www.copyright.com. Springer Publishing Company, LLC 11 West 42nd Street, New York, NY 10036 www.springerpub.com connect.springerpub.com/ Acquisitions Editor: Beth Barry Compositor: Transforma ISBN: 978-0-8261-3692-3 ebook ISBN: 978-0-8261-3693-0 DOI: 10.1891/9780826136930

20 21 22 23 24 / 5 4 3 2 1 Medicine is an ever-changing science. Research and clinical experience are continually expanding our knowledge, in particular our understanding of proper treatment and drug therapy. The authors, editors, and publisher have made every effort to ensure that all information in this book is in accordance with the state of knowledge at the time of production of the book. Nevertheless, the authors, editors, and publisher are not responsible for any errors or omissions or for any consequence from application of the information in this book and make no warranty, expressed or implied, with respect to the content of this publication. Every reader should examine carefully the package inserts accompanying each drug and should carefully check whether the dosage schedules therein or the contraindications stated by the manufacturer differ from the statements made in this book. Such examination is particularly important with drugs that are either rarely used or have been newly released on the market. Library of Congress Cataloging-in-Publication Data Names: Schuele, Stephan U., editor. Title: A practical approach to stereo EEG / editor, Stephan U. Schuele. Description: First Springer Publishing edition. | New York, NY : Springer Publishing Company, 2021. | Includes bibliographical references and index. Identifiers: LCCN 2020038905 (print) | LCCN 2020038906 (ebook) | ISBN 9780826136923 (paperback) | ISBN 9780826136930 (ebook) Subjects: MESH: Electroencephalography--methods | Epilepsy--surgery | Stereotaxic Techniques | Brain Mapping--methods Classification: LCC RC386.6.E43 (print) | LCC RC386.6.E43 (ebook) | NLM WL 368 | DDC 616.8/047547--dc23 LC record available at https://lccn.loc.gov/2020038905 LC ebook record available at https://lccn.loc.gov/2020038906 Contact us to receive discount rates on bulk purchases. We can also customize our books to meet your needs. For more information please contact: [email protected]

Publisher’s Note: New and used products purchased from third-party sellers are not guaranteed for quality, authenticity, or access to any included digital components. Printed in the United States of America.

This book is dedicated to Adriana, Emilia, and Camilo who make it all worthwhile.

Contents

Contributors  xi Foreword  Jean Gotman, PhD  xvii Preface  xxi Abbreviations  xxiii

SECTION I.  INTRODUCTION 1. The History and Principles of Stereo EEG  3 Patrick Chauvel

SECTION II.  NONINVASIVE EVALUATION 2. Phase I Evaluation  15 Allyson A. Pickard and Christopher T. Skidmore 3. Advanced MRI Imaging  27 Eliane Kobayashi 4. Electromagnetic Source Imaging for Stereo EEG Planning  35 Richard C. Burgess and Rafeed Alkawadri 5. Nuclear Imaging  49 David B. Burkholder, Elson L. So, Benjamin H. Brinkmann, and Lily C. Wong-Kisiel 6. Patient Selection for Stereo EEG  61 Jessica W. Templer and Stephan U. Schuele

Case presentations for Chapters 13, 15, 17, and 18 are available by scanning the QR code found at the beginning of the book or following this link to Springer Publishing Company ConnectTM: http://connect.springerpub.com/content/ book/978-0-8261-3693-0

viii   | Contents

SECTION III.  TECHNICAL ASPECTS 7. Electrodes  73 Saurabh R. Sinha 8. Coregistration of Multimodal Imaging  79 Thandar Aung, Irene Z. Wang, and Andreas V. Alexopoulos 9. Invasive Monitoring  93 Elisabeth Landré and Francine Chassoux

SECTION IV.  ELECTRODE PLACEMENT 10. The Epileptogenic Zone: A Critical Reconstruction  105 Patrick Chauvel 11. Toward a Revised Conceptual Framework of Stereo EEG Anatomo-Clinico-Electrical Correlations  119 Philippe Ryvlin 12. MRI-Negative Versus MRI-Positive Epilepsy in the Context of Stereo EEG  123 Manuela Ochoa-Urrea, Stephen A. Thompson, and Samden Lhatoo

SECTION V.  PRACTICAL APPROACH 13. Temporal Lobe Epilepsy  135 Giridhar P. Kalamangalam and Sotiris Mitropanopoulos Includes case discussion presented by Giridhar P. Kalamangalam and Jean Cibula available on Springer Publishing Connect 14. Frontal Lobe Epilepsy  157 Francesca Bonini and Aileen McGonigal 15. Posterior Cortex Epilepsy  177 Fábio A. Nascimento, George Culler, Stephan U. Schuele, and Jay R. Gavvala Includes case discussion presented by George Culler and Stephan U. Schuele available on Springer Publishing Connect 16. Temporal Plus Epilepsy  193 Carmen Barba and Matteo Lenge 17. Operculo-Insular Epilepsy  201 Raluca Pana and Dang Khoa Nguyen Includes case discussion presented by Raluca Pana and Dang Khoa Nguyen available on Springer Publishing Connect

Contents   | ix 

18. Tuberous Sclerosis Complex  217 Rohini K. Coorg and Jurriaan M. Peters Includes case discussion presented by Rohini K. Coorg and Jurriaan M. Peters available on Springer Publishing Connect 19. Stereo EEG in Epilepsy Associated With Nodular Heterotopia  235 Hui Ming Khoo and François Dubeau

SECTION VI.  SURGICAL PLACEMENT 20. Stereo EEG Electrode Implantation Using a Stereotactic Frame  253 Joshua M. Rosenow 21. Robotic Placement of Stereo EEG Electrodes  265 Arman Jahangiri and Robert E. Gross 22. Stereo EEG Electrode Placement in Children  281 Vamsidhar Chavakula, Joseph R. Madsen, and Scellig Stone 23. Surgical Complications and Challenges in Stereo EEG  293 Hussein A. Zeineddine and Nitin Tandon

SECTION VII.  INTERPRETATION 24. Physiological Activity Recorded With Intracranial EEG: From Wakefulness to Sleep  303 Laure Peter-Derex, Nicolás von Ellenrieder, and Birgit Frauscher 25. Interpretation of Interictal Epileptiform Discharges  315 Shasha Wu, Naoum P. Issa, Sandra Rose, and James X. Tao 26. High-Frequency Oscillations  329 Karina A. González-Otárula and Julia Jacobs 27. Ictal Activity  339 Fabrice Bartolomei, Julia Scholly, and Stanislas Lagarde

SECTION VIII.  ELECTRICAL STIMULATION MAPPING 28. Technical Aspects of Electrical Stimulation Mapping  353 Jay R. Gavvala and Stephan U. Schuele 29. Electrical Stimulation for Functional Mapping During Stereo EEG Exploration  361 Agnès Trébuchon 30. Stimulation-Induced Seizures to Define the Epileptogenic Zone  383 Aileen McGonigal

x   | Contents

SECTION IX.  SURGICAL PROCEDURES 31. Laser Ablation for Epilepsy  399 Michael R. Jones and S. Kathleen Bandt 32. Stereo EEG-Guided Resections: Method and Technique  409 Jorge Alvaro Gonzalez-Martinez, and Zachary C. Gersey

SECTION X.  OUTCOME 33. Seizure Outcome After Stereo EEG Implantations  423 Lara Jehi 34. Promise and Perils of Stereo EEG-Guided Resections for  Cognition and Emotion  431 Daniel L. Drane, Nigel P. Pedersen, Kelsey C. Hewitt, and Taylor Jordan Index  439

Contributors

Andreas V. Alexopoulos, MD, MPH  Staff Physician, Epilepsy Center, Neurological Institute, Cleveland Clinic, Cleveland, Ohio Rafeed Alkawadri, MD  Associate Professor, Department of Neurology; Director, Human Brain Mapping Program, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania Thandar Aung, MD  Assistant Professor, Department of Neurology, Barrow Neurological Institute, Phoenix, Arizona S. Kathleen Bandt, MD  Assistant Professor, Department of Neurosurgery, Feinberg School of Medicine, Northwestern University, Chicago, Illinois Carmen Barba, MD, PhD  Neurologist, Pediatric Neurology Unit, Neuroscience Department, Meyer Children’s Hospital, University of Florence, Florence, Italy Fabrice Bartolomei, MD, PhD  Neurologist, Institut de Neurosciences des Systèmes, INSERM, Aix-Marseille University; Professor of Neurology, Epileptology Department, Timone Hospital, AP-HM, Marseille, France Francesca Bonini, MD, PhD  Lecturer, Hospital Practitioner, Institut de Neurosciences des Systèmes, INSERM, Aix-Marseille University; Epileptology Department, Timone Hospital, AP-HM, Marseille, France Benjamin H. Brinkmann, PhD  Associate Professor of Neurology, Assistant Professor of Biomedical Engineering, Mayo Clinic Alix School of Medicine, Mayo Clinic, Rochester, Minnesota Richard C. Burgess, MD, PhD  Staff Physician, Epilepsy Center, Neurological Institute; Professor, Department of Neurology, Cleveland Clinic, Cleveland, Ohio David B. Burkholder, MD  Assistant Professor, Department of Neurology, Mayo Clinic Alix School of Medicine, Mayo Clinic, Rochester, Minnesota

xii   | Contributors

Francine Chassoux, MD  Neurologist, Neurosurgery Department, GHU-Paris, Paris, France Patrick Chauvel, MD  Neurologist, Epilepsy Center, Neurological Institute, Cleveland Clinical Lerner College of Medicine, Case Western Reserve University, Cleveland, Ohio Vamsidhar Chavakula, MD  Neurosurgery Resident, Department of Neurosurgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts Rohini K. Coorg, MD  Assistant Professor, Division of Neurology, Department of Pediatrics, Texas Children’s Hospital, Baylor College of Medicine, Houston, Texas George Culler, MD  Epilepsy Fellow, Department of Neurology, Northwestern University, Chicago, Illinois François Dubeau, MD  Associate Professor of Clinical Neurology, Department of Neurology and Neurosurgery, Montreal Neurological Institute and Hospital, McGill University, Montreal, Quebec, Canada Daniel L. Drane, PhD  Associate Professor, Departments of Neurology and Pediatrics, Emory University School of Medicine, Atlanta, Georgia; Affiliate Associate Professor, Department of Neurology, University of Washington School of Medicine, Seattle, Washington Nicolás von Ellenrieder, PhD  Research Associate, Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada Birgit Frauscher, MD, PD  Neurologist, Associate Professor of Neurology, Analytical Neurophysiology Lab, Montreal Neurological Institute and Hospital, McGill University, Montreal, Quebec, Canada Jay R. Gavvala, MD, MSCI  Assistant Professor, Department of Neurology – Neurophysiology, Baylor College of Medicine, Houston, Texas Zachary C. Gersey, MD, MS  Resident, Epilepsy Center, University of Pittsburgh, Pittsburgh, Pennsylvania Jorge Alvaro Gonzalez-Martinez, MD, PhD, FAANS  Professor, Department of Neurological Surgery; Co-Director, Epilepsy Center, University of Pittsburgh, Pittsburgh, Pennsylvania Karina A. González-Otárula, MD  Neurology Resident, Department of Neurology, The University of Iowa, Iowa City, Iowa Robert E. Gross, MD, PhD  MBNA Bowman Professor, Departments of Neurosurgery and Neurology, Emory University School of Medicine, Atlanta, Georgia Kelsey C. Hewitt, PsyD  Postdoctoral Fellow, Department of Neurology, Emory University School of Medicine, Atlanta, Georgia Naoum P. Issa, MD, PhD  Assistant Professor, Department of Neurology, The University of Chicago, Chicago, Illinois

Contributors   | xiii 

Julia Jacobs, MD, MSc, PD  Associate Professor, Director of Pediatric Epilepsy, Department of Pediatric Neurology, Alberta Children’s Hospital, University of Calgary, Calgary, Alberta, Canada Arman Jahangiri, MD, PhD  Neurosurgery Resident, Department of Neurosurgery, Emory University School of Medicine, Atlanta, Georgia Lara Jehi, MD  Chief Research Information Officer, Professor of Neurology, Cleveland Clinic, Cleveland, Ohio Michael R. Jones, MD  Resident, Department of Neurosurgery, Northwestern University, Chicago, Illinois Taylor Jordan, BA  Postgraduate Student, Department of Neurology, Emory University School of Medicine, Atlanta, Georgia Giridhar P. Kalamangalam, MD, DPhil  Wilder Family Professor, Department of Neurology, University of Florida, Gainesville, Florida Hui Ming Khoo, MD, PhD  Assistant Professor, Department of Neurosurgery, Osaka University; Neurosurgeon and Epileptologist, Epilepsy Center, Osaka Unversity Hospital, Suita, Osaka Prefecture, Japan Eliane Kobayashi, MD, PhD  Assistant Professor, Faculty of Medicine, Department of Neurology and Neurosurgery, McGill University; Neurologist, Epilepsy Service and EEG Department, Montreal Neurological Institute and Hospital, Montreal, Quebec, Canada Stanislas Lagarde, MD, MsC  Neurologist, Institut de Neurosciences des Systèmes, INSERM, Aix-Marseille University; Epileptology Department, Timone Hospital, AP-HM, Marseille, France Elisabeth Landré, MD  Neurologist, Neurosurgery Department, GHU-Paris, Paris, France Matteo Lenge, MD, PhD  Engineer, Neuroscience and Neurosurgery Department, Children’s Hospital Meyer, Florence, Italy Samden Lhatoo, MD, FRCP  Epileptologist, John P. and Kathrine G. McGovern Distinguished University Professor, Department of Neurology, Houston McGovern Medical School, University of Texas, Houston, Texas Joseph R. Madsen, MD  Director, Epilepsy Surgery; Associate Professor, Department of Neurosurgery, Boston Children’s Hospital, Harvard Medical School, Boston, Massachusetts Aileen McGonigal, MD, PhD  Neurologist, Institut de Neurosciences des Systèmes, INSERM, Aix-Marseille University; Epileptology Department, Timone Hospital, AP-HM, Marseille, France

xiv   | Contributors

Sotiris Mitropanopoulos, MD  Assistant Professor, Department of Neurology, University of Florida, Gainesville, Florida Fábio A. Nascimento, MD  Neurology Resident, Department of Neurology, Baylor College of Medicine, Houston, Texas Dang Khoa Nguyen MD, PhD, FRCPC  Professor, Department of Neuroscience, University of Montreal; Neurologist and Epileptologist, Centre Hospitalier de l’Université de Montreal, Montreal, Quebec, Canada Manuela Ochoa-Urrea, MD  Postdoctoral Research Fellow, Department of Neurology, Houston McGovern Medical School, University of Texas, Houston, Texas Raluca Pana, MD, FRCPC  Assistant Professor, Department of Neurology and Neurosurgery, McGill University; Neurologist and Epileptologist, Epilepsy and EEG Department, Montreal Neurological Institute and Hospital, Montreal, Quebec, Canada Nigel P. Pedersen, MBBS  Assistant Professor, Department of Neurology, Emory University School of Medicine, Atlanta, Georgia Laure Peter-Derex, MD, PhD  Assistant Professor, Center for Sleep Medicine and Respiratory Diseases, Lyon University Hospital, Lyon Neuroscience Research Center, Lyon, France Jurriaan M. Peters, MD, PhD  Assistant Professor, Division of Epilepsy and Clinical Neurophysiology, Boston Children’s Hospital, Harvard Medical School, Boston, Massachusetts Allyson A. Pickard, MD  Clinical Assistant Professor, Department of Neurology, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania Sandra Rose, MD  Assistant Professor, Department of Neurology, The University of Chicago, Chicago, Illinois Joshua M. Rosenow, MD, FAANS, FACS  Director of Functional Neurosurgery, Professor of Neurosurgery, Department of Neurology and Physical Medicine and Rehabilitation, Feinberg School of Medicine, Northwestern University, Chicago, Illinois Philippe Ryvlin, MD, PhD  Professor, Department of Clinical Neuroscience, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland Julia Scholly, MD, PhD  Neurologist, Institut de Neurosciences des Systèmes, INSERM, Aix-Marseille University; Epileptology Department, Timone Hospital, AP-HM, Marseille, France Stephan U. Schuele, MD, MPH  Professor of Neurology, Physical Medicine and Rehabilitation, Feinberg School of Medicine, Northwestern University; Epilepsy Section Head, Department of Neurology, Medical Director, Neurological Testing Center, Northwestern Memorial Hospital, Chicago, Illinois

Contributors   | xv 

Saurabh R. Sinha, MD, PhD  Associate Professor, Department of Neurology, Duke University, Durham, North Carolina Christopher T. Skidmore, MD  Associate Professor, Department of Neurology, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania Elson L. So, MD  Professor, Department of Neurology, Mayo Clinic Alix School of Medicine, Mayo Clinic, Rochester, Minnesota Scellig Stone, MD, PhD, FRCSC  Director, Stereotactic and Functionary Neurosurgery; Assistant Professor of Neurosurgery, Boston Children’s Hospital, Harvard Medical School, Boston, Massachusetts Nitin Tandon, MD  Professor, Vice-Chairman, Department of Neurosurgery, Co-Director Texas Institute of Restorative Neurotechnologies, University of Texas Health Science Center at Houston, Houston, Texas James X. Tao, MD, PhD  Associate Professor, Department of Neurology, The University of Chicago, Chicago, Illinois Jessica W. Templer, MD  Assistant Professor, Department of Neurology, Feinberg School of Medicine, Northwestern University, Chicago, Illinois Stephen A. Thompson, MD  Epileptologist, Assistant Professor, Department of Neurology, Houston McGovern Medical School, University of Texas, Houston, Texas Agnès Trébuchon, MD, PhD  Neurologist, Institut de Neurosciences des Systèmes, INSERM, Aix-Marseille University; Professor of Neurology, Clinical Neurophysiology Department, Timone Hospital, AP-HM, Marseille, France Irene Z. Wang, PhD  Staff Scientist, Epilepsy Center, Cleveland Clinical Neurological Institute, Cleveland, Ohio Lily C. Wong-Kisiel, MD  Associate Professor of Neurology and Pediatrics, Department of Neurology, Mayo Clinic Alix School of Medicine, Mayo Clinic, Rochester, Minnesota Shasha Wu, MD, PhD  Associate Professor, Department of Neurology, The University of Chicago, Chicago, Illinois Hussein A. Zeineddine, MD  Resident, Department of Neurosurgery, University of Texas Health Science Center at Houston, Houston, Texas

Foreword

The surgical treatment of medically refractory epilepsy can bring a cure to a high proportion of patients. This treatment is complex, requiring the expertise of a large multidisciplinary team including neurologists, neurophysiologists, imaging specialists, neuropsychologists, and neurosurgeons. The multidisciplinary investigation led by these specialists can be performed by noninvasive methods such as EEG and MRI, and in many cases, it leads to a clear surgical target, that is, a piece of brain tissue that can safely be removed in an operation likely to lead to seizure freedom and minimal functional deficit. In an important proportion of patients, however, such a noninvasive investigation leads to uncertain or ambiguous results. An invasive approach may then be considered. The most common is to place electrodes intracranially, thus bypassing the major barrier created by the skull, which blurs and dramatically attenuates the electrical signals generated by the brain (even though magnetoencephalography is not affected by the skull, it is quite similar to EEG with respect to seeing deep or small brain generators). In addition, intracranial electrodes can be placed in deep brain structure, hopefully in proximity to the sources of epileptic discharges; they are a unique way to provide critical information regarding the brain region where seizures start. The placement of intracranial electrodes is a surgical procedure not without risk and it must therefore be considered only if it has a good chance of success. The major weakness of intracranial electrodes is that they are limited to record from a small fraction of the brain: the “field of vision” of each electrode contact is a few millimeters around itself, and even if one puts 100 or 200 electrode contacts, only a small fraction of brain volume can be investigated. The main challenge of intracranial investigations is therefore to preselect targets that have a high probability of being in the region where seizures start. During the past decades, two main approaches to intracranial EEG investigations have been developed: the first is the stereotaxic placement of needle-like electrodes, going inside the brain at locations defined in a three-dimensional coordinate system, a method pioneered in France by Talairach and Bancaud in the 1960s and 1970s and called stereoelectroencephalography (SEEG). This method has been used in many centers, mainly in France, Italy, and Canada. A second approach, pioneered by Lüders in the 1980s, is the placement of flexible subdural electrode strips and grids, lying on the cortical surface. This approach has been widely adopted in the United States and in some other countries. Many symposia have discussed the merits of the two approaches over the years, with the respective proponents of each method usually concluding that their method was the best. In the last few years, however, the tide has turned and many of the users of subdural grid electrodes are adopting SEEG, particularly in the United States, in part because of the clear difference in the rate of complications and the relative simplicity of the implantation process when using robotic assistance, and in

xviii   | Foreword

part because of a better understanding of the SEEG approach. That is why this book, edited by Stephan Schuele, comes at such an opportune time. It is edited and written in large part by people from the United States who have witnessed this change and are in an excellent position to present the SEEG method in the context of a community accustomed to subdural grid electrodes. Many chapters are also written by long-time SEEG specialists: the SEEG approach is introduced by Dr. Patrick Chauvel, the master of the French school who is actively spreading the word around the world. Others who can be mentioned include Drs. Bartolomei, Chassoux, Ryvlin, and Dubeau. The book is a complete and practical guide to thinking and doing SEEG which will be a solid reference to practitioners around the world. It starts with a review of the main noninvasive preimplantation investigations and how they inform the selection of patients and the implantation scheme. It proceeds to implantation techniques and issues specific to the different brain regions, including the complex question of temporal-plus epilepsies and investigation of regions that were practically impossible to reach with subdural electrodes such as the insula. The problem is then attacked according to the different types of epileptogenic lesions; as examples, the tuberous sclerosis complex and periventricular nodular heterotopia can be investigated only with SEEG. The critical importance given by the French and Italian epilepsy schools to the relation between a very carefully observed clinical symptomatology and functional anatomy is evident throughout this book. An SEEG exploration is more than any other investigation in this field a partnership between the neurologist and the neurosurgeon. The implantation scheme must consider not only the possible epileptogenic zone but also the constraints of implantation and what will be possible in an eventual surgery. The four chapters dealing with surgical issues cover well the questions of surgical risk (Tandon), implantation techniques based on stereotaxic frames (Rosenow), or robotics (Jahangiri and Gross), as well as the unique advantages and challenges of using SEEG in young children (Chavakula et al.). The subdural electrode approach is well suited for functional mapping of eloquent cortex, but SEEG can also be effectively used for functional mapping, giving good access to sulci and deep structures, as explained by Drs. Gavvala and Schuele and by Dr. Trébuchon; it is also an effective tool for triggering seizures, which can help confirm spontaneous seizure origin, as described by Dr. McGonigal. Although the main purpose of an SEEG implantation is to record seizures, the chapter by Dr. Wu et al. dedicated to interictal epileptiform explains how to interpret the complex interictal patterns, which in some cases can be more important than the seizures themselves. The chapter by Drs. Gonzalez-Otarula and Jacobs discusses the possible role of high frequency oscillations as another interictal biomarker of the epileptogenic zone. Dr. Peter-Derex’s chapter on physiologic activity also explains how it is now possible to interpret the SEEG background by comparison to normal intracerebral patterns of the different brain regions. The thinking SEEG mind-set is well illustrated in the chapter by Drs. Gonzalez-Martinez and Gersey on resection who discuss how to take into account the structural and functional anatomy, and the circuitry of the brain to conceptualize the epileptogenic zone, that is, to build a hypothesis that could fully explain the preimplantation observations, as well as prepare alternative hypotheses. The book concludes with the key issue: How successful is surgery in patients who undergo an SEEG investigation? The results presented by Dr. Jehi are generally encouraging, with more than half of the patients becoming seizure-free. This is half of a group of patients who would probably not be offered surgery without the opportunity of an SEEG study. Of course, results vary a lot from study to study, probably in relation to the heterogeneity of patient populations and to the varying levels of expertise. For this last factor, I am confident that this book will contribute to a widespread increase of understanding of how to approach the complex but often rewarding SEEG investigation. It will be a useful reference for the attendees of the several intensive workshops being given around the world where the art and the science of SEEG are being taught in a realistic clinical context, also helping to raise the level of expertise and the awareness of the real-life intricacies of complex epilepsy surgery.

Foreword   | xix 

To conclude, one word of caution, also given by many contributors to the book, and independent of the type of electrode used: when interpreting intracranial EEG results, you see only what happens where there are electrodes, and most of the brain has no electrodes. What looks like the onset of a seizure can result from propagation from an unexplored region. Furthermore, the region where seizures start may be a subset of the regions where seizures can start (how can we explain seizures stopping for one year after surgery but restarting with unchanged symptomatology?) There is yet much to understand about seizures and the brain. This book will go a long way in helping to treat today’s patients and also in leading to discoveries that will help even more effectively tomorrow’s patients. Jean Gotman, PhD Associate Professor Department of Neurology and Neurosurgery Montreal Neurological Institute McGill University Montreal, Quebec, Canada

Preface

Over the last decade, stereo EEG has become the predominant method across the world to invasively explore patients with focal epilepsy who are potential candidates for resective surgery. This shift away from open surgery with subdural grid implantation required many epilepsy centers to introduce major workflow adaptations, investment in surgical and imaging technologies, and above all, seek training in placement and interpretation of depth electrodes recordings. It became evident that despite the long tradition of stereo EEG, a comprehensive, practical textbook outlining the different steps and nuances of the methodology was missing. This book is covering all practical aspects of stereo EEG and will be a quintessential staple for anybody learning and working in the field of epilepsy surgery, including adult and pediatric epileptologists and neurophysiologists, functional neurosurgeons, technologists, and trainees in these areas. Almost all chapters feature illustrative cases to explain specific aspects and key concepts of the SEEG methodology, making it an excellent learning tool and enhancing its practical applicability. The section covering the practical approach to specific epilepsy syndromes includes voice-over slide presentations demonstrating the process of a systematic patient discussion, hypothesis generation, and electrode planning followed by data interpretation and delineation of the surgical resection. The book starts with the historical background and principles of stereo EEG followed by a detailed discussion of the role of the noninvasive evaluation including video EEG, imaging, and patient selection. The section on technical aspects includes a discussion of electrodes, multimodal data coregistration, and guidelines for invasive monitoring. After a part covering the conceptual framework of stereo EEG, a series of chapters discusses various anatomically and pathologically defined epilepsy syndromes with multiple case examples explaining the implantation strategy. The second part of the book goes over surgical aspects of stereo EEG electrode placement covering robotic and frame-based approaches, specific pediatric aspects, and potential complications. This is followed by a section on data interpretation of physiologic, interictal, and ictal epileptic activity. The conceptual and methodological aspects of electrical stimulation mapping are outlined in the following part. The book ends with discussing surgical procedures to remove the epileptogenic zone and a review of seizure and cognitive outcome with stereo EEG. It is my hope that the book provides a comprehensive source with all the essential clinical information needed to offer epilepsy surgery guided by stereo EEG. This first edition is

xxii   | Preface

a milestone in the field and will allow us to advance our understanding of the epileptogenic zone and to recognize areas in need for refinement of our methodological approach. I want to express my wholehearted gratitude to all the authors who provided their invaluable time and expertise to accomplish this feat and whose companionship while writing the book carried me through these difficult times of a global pandemic. Stephan U. Schuele, MD, MPH

Abbreviations

3D three-dimensional AFP-Oper anterior frontoparietal operculum A-ins anterior insula aBTLA anterior part of the basotemporal language area AC anterior commissure ACC anterior cingulate cortex AD afterdischarge AEC anatomo-electro-clinical AED antiepileptic drug AF arcuate fascicle AH anterior hippocampus Am amygdala aMT anterior mesial temporal ANG angular gyrus angs angular sulcus Ant Insula anterior insula APC anterior parietal cortex BA Brodmann area BESA brain electrical source analysis BNT Boston Naming Test BOLD blood oxygenation level-dependent BPNH bilateral periventricular nodular heterotopia BRW Brown–Roberts–Wells BTLA basotemporal language area cings cingulate sulcus CRW Cosman–Roberts–Wells cs central sulcus CS collateral sulcus CSF cerebrospinal fluid cSMA caudal supplementary motor area CTA CT angiogram Ctx cortex DA dimension of activation DC direct current

xxiv   | Abbreviations

DES direct electrical stimulation dLPFC dorsolateral prefrontal cortex DRE drug-resistant epilepsy DTI diffusion tensor imaging EA epileptic activity EC entorhinal cortex ECG electrocardiography ECoG electrocorticography EDH epidural hematoma EI Epileptogenicity Index EM epileptogenicity map EMG electromyogram EMU epilepsy-monitoring unit EOG electrooculogram EP evoked potential ER energy ratio ESI electric source imaging ETLE extratemporal lobe epilepsy EZ epileptogenic zone EZF epileptogenic zone fingerprint EZN epileptogenic zone network F Op pars opercularis of the inferior frontal gyrus F-Oper frontal operculum FCD focal cortical dysplasia FCD 2 focal cortical dysplasia type 2 FDG fluorodeoxyglucose FEF frontal eye field FLAIR fluid-attenuated inversion recovery FLE frontal lobe epilepsy FLNA Filamin A FLS Frontal lobe seizure fMRI functional MRI FOP frontal operculum Fr frontal FR fast ripple Fus fusiform gyrus GABA gamma aminobutyric acid GL gyrus lingualis GM gray matter GSM gyrus supramarginalis GWB gray and white matter junction blurring HARNESS Harmonized Neuroimaging of Epilepsy Structural Sequences Hc hippocampus HD-EEG high-density EEG He heterotopia HF high frequency HFA high-frequency ictal activity HFO high-frequency oscillation HFS high-frequency stimulation Hip hippocampus Hipp ant anterior hippocampus HRAIM Head Ring Assembly With Intubation Mounts

Abbreviations   | xxv 

HRD head ring drive HS hippocampal sclerosis ICEEG intracranial EEG ICH intracerebral hemorrhage IE insular epilepsy IED interictal epileptiform discharge IFG inferior frontal gyrus ILAE International League Against Epilepsy Inf Fr inferior frontal gyrus INS ant anterior insula INS postposterior insula IO insulo-opercular IOZ ictal onset zone IPL inferior parietal lobule IPS intraparietal sulcus ITS inferior temporal sulcus IV intravenous IZ irritative zone L left LA lesion, anterior LAST left anterior superior temporal lf lateral fissure LF low frequency LFS low-frequency stimulation LITT laser interstitial thermal therapy LLIN left lingula LM lesion, middle LMC left middle cingulate LMST left middle superior temporal LOC left occipital LPC left posterior cingulate LPFC lateral prefrontal cortex LPIN left posterior insula LPST left posterior superior temporal LSF left superior frontal LVFA low-voltage fast activity LVFD low-voltage fast discharge LVRD low-voltage rapid discharges MAP morphometric analysis program Mc magnetoencephalography cluster MCC midcingulate cortex MEG magnetoencephalography MFG middle frontal gyrus MIFE medically intractable focal epilepsy MLF middle longitudinal fascicle MNE-fc minimum-norm estimate of a narrow band MNI Montreal Neurological Institute MPFC medial prefrontal cortex MR magnetic resonance MRE medically refractory epilepsy MRI-SLA MRI-guided stereotactic laser ablation MRLITT magnetic resonance-guided laser interstitial thermal therapy

xxvi   | Abbreviations

MSI magnetic source imaging MTG middle temporal gyrus MTLE mesial temporal lobe epilepsy mTOR mechanistic target of rapamycin MTS mesial temporal sclerosis MUSIC Multiple Signal Classification NH nodular heterotopia NIN noninvolved network NMDA N-methyl-D-aspartate NPO nil per os NREM nonREM NS neurostimulation NTLE neocortical temporal lobe epilepsy Oc occipital cortex OF fronto-opercular OFC orbitofrontal cortex OP parietal operculum OR operating room PACU postanesthesia care unit Pars Operc pars opercularis Pars Orb pars orbitalis Pars Triang pars triangularis pBTLA posterior part of the basotemporal language area PC posterior commissure PCC posterior cingulate cortex PCG postcentral gyrus PCUN precuneus PDS paroxysmal depolarization shift PFC prefrontal cortex PGO ponto-geniculo-occipital PIN posterior insular pis primary intermediate sulcus of Jenssen planum T planum temporale PM premotor PMG polymicrogyria pMTL posterior mesial temporal PNH periventricular nodular heterotopia poc poscentral sulcus pos parieto-occipital sulcus Post-C postcentral POSTS positive occipital sharp transients of sleep PP planum polare PPC posterior parietal cortex Pre-C precentral pSTG posterior part of the superior temporal gyrus PT planum temporale PVH posterior periventricular PZ propagation zone PZN propagation zone network R right RA right amygdala RAH right anterior hippocampus

Abbreviations   | xxvii 

RAISG right anterior insular superior gyrus RANT right anterior temporal RBT right basal temporal RCT randomized controlled trial RFA radiofrequency ablation RFTC radiofrequency thermocoagulation RIA radioimmunoassay RIA right inferior anterior RPS right posterior superior RNS responsive neurostimulation ROF right orbitofrontal ROSA Robotic Stereotactic Assistance RIFG right inferior frontal gyrus RMFGA right middle frontal gyrus RPILG right posterior insular lower gyrus RPH right posterior hippocampus RPIN right posterior insular RRS right retrosplenium RSA rhythmic slow activity rsfMRI resting state functional MRI rSMA rostral supplementary motor area RTP right temporal pole SAH subarachnoid hemorrhage SD-ECoG subdural electrocorticography SDE subdural electrode SDG subdural grid SDH subdural hematoma SECD single equivalent current dipole SEEG stereoelectroencephalography SFG superior frontal gyrus SFS superior frontal sulcus SISCOM subtraction ictal single-photon emission computed tomography coregistered to MRI SLA stereotactic laser ablation SLAH stereotactic laser amygdalohippocampotomy SLF superior longitudinal fascicle sLORETA standardized low-resolution brain electromagnetic tomography SMA supplementary motor area SMG supramarginal gyrus SNH subcortical nodular heterotopia SOP seizure onset pattern SOZ seizure onset zone SPECT single-photon emission computed tomography SPL superior parietal lobule SPM statistical parametric mapping sps subparietal sulcus SQUID superconducting quantum interference device STATISCOM statistical ictal SPECT coregistered to MRI STG superior temporal gyrus STGa superior temporal gyrus A sTPJ temporo parietal junction STS superior temporal sulcus

xxviii   | Abbreviations

STS ant anterior aspect of the superior temporal sulcus STW saw-tooth waves SUDEP sudden unexpected death in epilepsy Supra Marg supramarginal T1 ant anterior aspect of the superior temporal gyrus T2 ant anterior aspect of the middle temporal gyrus T3 ant anterior aspect of the inferior temporal gyrus T1-WI T1-weighted image TA2 temporal region A2 Tc99m-ECD technetium-99m ethyl cysteinate dimer TL temporal lobe TLE temporal lobe epilepsy TMS transcranial magnetic stimulation Tp temporal pole TP temporo-parietal TPE temporal plus epilepsy TPO temporo-parieto-occipital TSC tuberous sclerosis complex UCHRA Universal Compact Head Ring Assembly VBM voxel-based morphometry VCA line vertical line passing through anterior commissure vLPFC ventrolateral prefrontal cortex WM white matter XTLE extratemporal lobe epilepsy

Section I. Introduction

1 The History and Principles of Stereo EEG Patrick Chauvel

THE PRESTEREO EEG ERA Stereoelectroencephalography (SEEG) was invented in the late 50s by Jean Talairach, a neurosurgeon, and Jean Bancaud, a neuropsychiatrist and electroencephalographer.1 This method was based upon recording of brain electrical activity by intracerebral electrodes, stereotactically implanted in preidentified cortical and subcortical structures. When the technique was developed only electrocorticography (ECoG) was utilized. ECoG was an interictal investigation that did not provide any information on actual seizure onset. Furthermore, ECoG had to be carried out after craniotomy, so the decision of whether and how to operate had to be made at the same time in the operating room. In contrast, SEEG offered a method that allowed the investigative presurgical and therapeutic surgical stages to be separated. Using data from ictal and interictal EEG recordings, complemented with the results of local electrical stimulation applied through adjacent contacts of the implanted electrodes, surgical planning could be rationally prepared.2 The impetus for Bancaud and Talairach to start an epilepsy surgery program can be traced back to Henry Hécaen, a neuropsychologist who visited the Montreal Neurological Institute (MNI) and examined postsurgical patients with Wilder Penfield and Brenda Milner. After a year of study in 1952, Hécaen returned to Hospital Sainte-Anne, Paris, full of enthusiasm for surgical treatment of epilepsy. He convinced Gabriel Mazars, a neurosurgeon, that he should not wait to start an epilepsy surgery program based on the MNI method. At that time, Jean Talairach worked in the Department of Neurosurgery at the same hospital with his mentor Marcel David.3 Since 1946 he had been developing stereotactic approaches to functional neurosurgery for chronic pain and movement disorders.4 Jean Bancaud was a pupil of Henri Fischgold and had presented a doctoral thesis on the relationship between neuropsychological deficits and EEG features in patients with cerebral tumors. Then Fischgold’s group joined David’s in Sainte-Anne. Bancaud and Talairach met. They were instantly immersed in an outstanding medical and scientific environment, they were Mr. and Mrs. Dell in neurophysiology, H. Hécaen and J. de Ajurriaguera in neuropsychology. Sainte-Anne was a psychiatric hospital founded in 1651. In the 50s, the department of psychiatry was dynamic and renowned as a pioneer group in biological psychiatry: Jean Delay, Pierre Deniker, with Henri Laborit demonstrated the “neuroleptic” effects of chlorpromazine in 1952. Hécaen had clearly a strong influence on Bancaud, who absorbed the lessons of Penfield on localization, and understood at a very early stage the complementary information provided

4   | SECTION I. INTRODUCTION

by clinical signs and symptoms and the EEG, whether due to lesions or occurring during seizures. This insight led him to look for methods more accurate than ECoG for localization.

LAYING THE FOUNDATION Meanwhile, Talairach was coming to a decisive turning point in his own methodological process. He had worked on the design of a surgical frame that would fit human brains of various sizes and would allow accurate repositioning. In 1947, the Talairach frame was born and improved over the years. Obsessed by increasing the accuracy of stereotaxic localization of deep structures, Talairach had also invented in 1949 the double grid system, a device made of two parallels grids attached to the stereotaxic frame, and through which locator needles or electrodes could be guided into the brain (Figure 1.1). The double grids served to minimize distortion due to X-ray diffraction in ventriculography images and thus offered the opportunity to precisely align radiological and postmortem anatomical data. Note that to further reduce distortion, Talairach later conceived a very large surgery room so that the X-ray tube was placed almost 5 m away from the grids.2 His vision was at that time very unusual. He regarded stereotactic methods as optimal for analysis of human brain anatomy in three-dimensional space, rather than as merely an instrument for reaching a target in the brain. Talairach built up a coordinate system based on the anterior commissure (AC) to posterior commissure (PC) base line and studied human neuroanatomy in reference to this line. This method also allowed the “normalization” of anatomical data from different brains such that data from the functional investigations of different patients could be displayed on the same summarizing chart. This enormous body of work led to the publication of two anatomical atlases. The stereotaxic frame, the double grids, and the AC–PC system were put together by Talairach to precisely map the anatomic stereotaxy of basal ganglia. In 1957 he published the first stereotaxic atlas of basal ganglia,5 a book that soon became a reference for neurophysiologists and neurosurgeons. It took Talairach 10 years more to extend the stereotaxic approach to the entire brain. For this, he implemented the concept of proportional scaling to account for differences in brain size in the three directions of the AC–PC system. Working with a rare tenacity, he validated the concept in tens of postmortem studies and published in 1967 a second stereotaxic atlas devoted to the telencephalon, which also became a reference and a basis for SEEG.6 Bancaud became more and more attracted to Talairach’s method, which promised to localize the sources of scalp EEG discharges. However, at that time, stereotaxy was oriented toward pain, dyskinesias, parkinsonian tremor, and some otherwise inoperable tumors. Working with M. B. Dell, Bancaud got progressively convinced of the clinical utility of epilepsy surgery but considered that methods of investigation current at that time were poorly adapted to epileptic phenomena.7 After Talairach’s completion of the 1957 anatomical atlas, Marcel David took Talairach’s views into account and supported a project to create an operating room dedicated to stereotactic surgery. This theater needed to be large enough to satisfy the physical requirements of teleradiography and allow parallel X-ray beams in order to avoid distortions of skull, vessels, ventricles, and, crucially, the frame and double grids used for guiding the placement of intracranial electrodes. The concept of a stereotactic surgery room was born, and such a suite was opened in Sainte-Anne in 1959 in the walls of an ancient chapel. Bancaud saw the potential of a spatial conceptualization of the human brain promoted by Talairach’s method and began working with him to develop applications of stereotactic functional exploration to presurgical investigations of intractable epilepsies. In the minds of Talairach and Bancaud, Penfield’s localization approach could be thoroughly applied only through electrophysiological recording of seizures directly from the involved brain structures. This goal was achieved in 1959 using Talairach’s stereotactic method to accurately place intracerebral electrodes with respect

1. The History and Principles of Stereo EEG   | 5 

FIGURE 1.1  The Talairach stereotaxic frame with the double grids allowing orthogonal and oblique

electrodes implantation (circa 1962).

to reliable anatomical markers and then record the electrical activity of multiple cerebral structures during the course of a seizure. Benefiting from this extraordinary technical achievement, Bancaud developed a careful analysis of anatomical–physiological–clinical correlations that could directly determine surgical strategy by utilizing electrophysiological recordings of seizures precisely localized in cerebral space. With these data, Bancaud studied the spatiotemporal dynamics of seizure discharges with respect to their clinical features. In particular, he first described the respective contributions of medial and lateral cortical areas in the organization of temporal and frontal seizures: the role of amygdala and hippocampus versus the temporal neocortex in temporal seizures (Figure 1.2),8,34 and the role of supplementary motor area and cingulate area 24 versus the dorsolateral, ventrolateral, and ventromedial frontal cortices.9 These studies had immediate repercussion on the practice of epilepsy surgery.

6   | SECTION I. INTRODUCTION

FIGURE 1.2  Stereo EEG recording in temporal lobe epilepsy circa 1962. Seizure onset with a fast ac-

tivity in left amygdala (NAG) and left hippocampus (CAG) spreading to the posterior parahippocampal gyrus (gyr pH post G). The two bottom traces are scalp EEG, that was simultaneously recorded.

CONCEPTUALIZATION OF STEREO EEG Intracerebral ictal recordings provided data so complex that their interpretation needed a strategic roadmap. The designations of the cortical areas according to their hierarchical involvement in the epileptogenic process required precise definitions. The region where seizures originate was obviously in the forefront of presurgical planning. The “epileptogenic zone” was differentiated from the space occupied by interictal activity, or the “irritative zone,” and from the space occupied by any morphological alteration or lesion supposed to be related to the epilepsy, or the “lesional zone.” Talairach and Bancaud showed that the lesional and irritative zones had a variable topographic relationship with the epileptogenic zone.10 Talairach and Bancaud developed a new type of epilepsy surgery, based on a three-dimensional representation of the epileptogenic, irritative, and lesional zones. With this novel method, the surgical plan could be carefully prepared and adapted to the individual patient’s case by referring radiological, physiological, and clinical data by physical matching to the same patient through the stereotaxic method. The epileptogenic and lesional zones' relationship depending on their cortical localization and on the different types of lesions underwent ongoing evaluation and validation. Today the use of neuroimaging has simplified the question of the lesional zone. At that time, only pneumoencephalography, lipiodol ventriculography, and cerebral angiography were available. Talairach looked for localizing and then operating responsible lesions with the help of localizing the epileptogenic zone. SEEG was so precise that he discovered numerous pseudotumoral, nonexpansive lesions (they would be identified later as dysembryoplastic neuroepithelial tumor and cortical malformations) that had not been previously shown by the neuroradiological techniques. He therefore encouraged Bancaud to find electrical criteria for detecting lesions. In SEEG terminology, the term “lesional zone” referred to the area occupied by various

1. The History and Principles of Stereo EEG   | 7 

types of slow-wave activity (characterized by waveform, frequency, and reactivity). The correlations were clearly superior to those given by the available neuroradiological images. Talairach and Bancaud, using the SEEG method, extended its application to map better the extent and anatomical location of astrocytomas or oligodendrocytomas. They correlated SEEG findings with the results of stereotaxic biopsies.11 According to histopathology, surgical resections avoiding functional deficits or targeting stereotactic interstitial radiotherapy could be preplanned. The topographic transitions between electrical silence inside glial tumors, slow waves, spiking, subclinical paroxysms, and seizure activity were analyzed. Beyond their clinical utility for tumor surgery, all this collection of anatomical-electrical data served as a model to interpret SEEG in cases of epilepsy. Definitions were less precise for the cortical areas producing interictal spikes and ictal discharges. When Bancaud began to record seizures using intracerebral electrodes, he noticed that the respective topography of the two types of discharge, interictal and ictal, were far from fully overlapping. He therefore used the terms "irritative zone" and "epileptogenic zone" to describe and differentiate the spatial extent of interictal spiking and ictal discharges respectively. The term epileptogenic zone thus referred to the anatomical structure(s) where primary organization of seizures took place.12 The role of fast synchronizing discharges that might involve more often distinct but interconnected regions than a single region was emphasized. Even though they could not give it a scientific content, the concept of epileptogenic zone was central to the development of SEEG. Bancaud and Talairach were immediately convinced to use it as the basis of their surgical strategy rather than the location of interictal discharges used by Penfield. Since the seizure was the symptom to be cured, it was the area of seizure organization that had to be determined. However, a strict correlation between the area(s) of seizure onset and emergence of clinical semiology was often difficult to establish in practice, because ictal signs and symptoms appeared and overlapped in time as the ictal discharge propagated to different cerebral structures. Understanding the spatiotemporal dynamics of the seizure as revealed by the sequence of signs allowed Bancaud to elaborate an extrapolation of its anatomical origin. This was a genial intuition. With this principle of extrapolation, the characterization of the initial symptom(s) or sign(s) appeared to be no more important than any later part of the seizure. Bancaud’s seizure patterns included order and sequence of semiological elements as their crucial features and were always referred to anatomy.13 Bancaud and Talairach definitely regarded a static view of seizure genesis represented by the gross topography of interictal and/or ictal onset abnormalities as insufficient for determining the area and volume of brain tissue to resect. They believed that a dynamic model of the discharge should be inferred from detailed consideration of all the data giving good discrimination between pacemakers versus active and passive relays of propagation. This was a prerequisite for surgical therapy in order to remove the smallest possible volume, and an intellectual discipline able to provide sound outcome information for further refinement of surgical criteria. With time the novel technique of Bancaud and Talairach evolved into a comprehensive and established method, where electrophysiological advances increased knowledge of anatomo-functional correlations giving rise to the beginnings of the functional exploration of networks.

TECHNICAL REQUIREMENTS Fixating double grids onto the stereotactic frame ensured safe implantation of intracerebral electrodes recording from lateral and medial areas of the cerebral cortex. Safety was further guaranteed by routine stereotactic angiography, allowing electrode placement that avoided vulnerable blood vessels. The practice of stereotactic angiography was associated with a very low risk. The stereotactic atlas of the telencephalon, containing frontal, axial, and sagittal brain slices referenced to the AC–PC baseline, was used as a guide to the topography of cortical

8   | SECTION I. INTRODUCTION

areas. The Talairach grid system based on AC–PC provided a statistical representation of the spatial distribution of cortical sulci. The accuracy of this approach was satisfactory for designing electrode implantations or for comparing patients but was insufficient for detailed interpretation of an individual patient’s SEEG data. A reconstruction of the precise position of the electrode contacts after implantation relative to the individual patient’s anatomy was still necessary. Fortunately, Gabor Szikla had taken advantage of the need for stereotactic angiography and had carefully described cerebral cortical blood vessels and their close relationships to the convolutions of gyri and sulci. He showed that cortical vessels mold themselves to the form of the gyri. He was able to extract gyral form and dimensions from a meticulous analysis of vascular anatomy. This discovery led to the concept that “vascular lamina” could provide constant landmarks for interpretation of anatomical variability among individuals and turned out to be a very precise tool for deducing the actual location of electrode contacts before the availability of modern imaging methods.14 The anatomical method of Talairach has not only survived the development of magnetic resonance-derived stereotactic techniques but continues to be the referential basis for SEEG.15

FROM COMPREHENSIVE BRAIN SAMPLING TO RATIONAL SURGERY SEEG offered high spatial and temporal resolution and a high power of localization. This was immediately evident in the early 1960s when the role of amygdala and hippocampus in temporal lobe seizures was discovered.16 During the next decade a major pitfall was to confuse stereotactic depth electrode explorations and real SEEG. This former type of implantation could be targeted for instance to preselected sites thought to be the source of paroxysmal activity on the basis of empirical arguments. This could be bilateral and symmetrical electrodes in mesial temporal structures only in cases of temporal lobe epilepsy. In contrast, SEEG is displaying the spatiotemporal dynamics of seizures in relation to the anatomy. This was contingent on using multiple multilead electrodes to give a distribution of spatial sampling sufficient to capture as much of the paroxysmal activity associated with the seizure as possible. A too compartmentalized or restricted view of ictal intracerebral activity would be insufficient to understand correlations with clinical semiology. The essential question of reliable spatial sampling was whether recording from too restricted a brain area would prevent SEEG from providing an accurate view of the electrical organization of the entire epileptogenic zone. An ideal distribution of sites for intracerebral recordings provides the neurophysiologist with an immediate view of the dynamics of ictal and interictal events. Physiological interpolation between the recording sites would thus depend on a correct strategy of implantation, which requires that the implantation be managed within a coherent method based on the individual patient’s symptoms. In this context, a strategy of electrode implantation demands a clear hypothetical framework regarding the distribution of the epileptogenic and lesional zones derived from a detailed analysis of the electrophysiological features of the interictal state, the video-EEG electrical–clinical correlations of the ictal state, and their relation to anatomical anomalies. SEEG was conceived for its capability to validate or invalidate a principal hypothesis, and at the same time to support or eliminate an alternative hypothesis. This “anatomo-electro-clinical” method produced types of data sets that may interact and validate one another. For example, the coincidence of clinical and electrical onset, or the precedence of electrical onset to clinical onset, both tend to confirm accurate positioning of a recording electrode. An indispensable tool is provided by intracerebral stimulation through stereotactically implanted electrodes, a technique proven reliable for showing the relationships between the different involved areas and the ictal semiology. Electrical stimulation can confirm topography of the ictal onset by provoking a habitual seizure and gives useful insight into the organization of the epileptogenic zone, by separating relay and subrelay areas essential

1. The History and Principles of Stereo EEG   | 9 

for building up individual ictal symptoms and signs, as well as their modes of clustering. Such data are helpful to the surgeon in planning a strategy of removal and/or disconnection. The ultimate objective of SEEG is to provide the surgeon with an integrated view of the epileptogenic process, based on the definition of the epileptogenic zone and its overlap with the lesional zone. Eventual matching of clinical, physiological, and anatomical data is essential, and the role of this stereotactic method in preoperative planning is exactly to achieve this objective. During surgery, anatomical and physiological data that describe the epilepsy and the function of related cortical areas are referenced to three-dimensional coordinates, which allows the surgeon to perform a comprehensively preplanned cortectomy within predetermined functional and vascular constraints.

THE STEREO EEG SCRIPTURES The development of SEEG ushered in a new era that coincided with a complete structural reorganization of the Sainte-Anne Department of Neurosurgery. Alain Bonis, a neurologist and pupil of Raymond Garcin, joined the team, completed by Gabor Szikla, an emigrant from Hungary in 1956, who was to devote his career to anatomy and stereotactic surgery under Jean Talairach. In 1960, David and Fischgold moved away to La Pitié-Salpétriere, and Talairach became the head of a new Department of Functional Neurosurgery at Hospital Sainte-Anne. This was a period of scientific productivity. The role of temporal limbic structures, the amygdala, and the hippocampus in clinical symptomatology and electroclinical organization of temporal lobe seizures was demonstrated. Clinical manifestations of absences that had been attributed to discharge of “centrencephalic” nuclei and as such considered as signs of generalized epilepsies by Penfield, Gastaut, and others could be elicited by stimulation of frontal anterior and intermediate areas, especially from their medial aspect.17 The anatomical-functional organization of the supplementary motor area was described as a part of the premotor systems and considered distinct from language areas.18 The first milestone in the establishment of SEEG fundamentals was the book La Stéréoelectroencéphalographie dans l’Épilepsie published in 1965.19 A modern and dynamic view of the focal epilepsies emerged from the observations, reflections, and discussions it collected. This book marked a break with the past, and, beyond the specific questions it raised in the field of epilepsy surgery, it provided an authoritative justification of the critical necessity for studying seizures when studying epilepsy. SEEG offered the opportunity of analyzing the dynamics of epileptic phenomena in humans, but also represented a remarkable tool for the study of normal neurophysiology. Pierre Buser, a prominent neurophysiologist interested in the cerebral cortex, undertook such studies in cooperation with Bancaud after 1964. He worked on motor systems, particularly the pyramidal system and sensory polymodal afferents to the frontal cortex. One of the models he used was the startle reaction in the anesthetized cat. This led to a better understanding of startle physiology, and the startle epilepsies in hemiplegic children being investigated by Bancaud.20 Buser influenced Bancaud’s work on various aspects of epilepsies with motor manifestations. Besides the startle epilepsies, they worked on the connections of the supplementary motor area,21 and together demonstrated the cortical origin of myoclonus and seizures in the Kojevnikov’s syndrome or epilepsia partialis continua,22 as well as the cortical origin of postanoxic action myoclonus (Lance and Adams syndrome).23 Buser used SEEG not only for functional mapping, but also to study corticocortical connections and their facilitation inside the epileptogenic zone.24 He was the first to use single pulse electrical stimulation to explore the excitability of corticocortical pathways and was in fact the pioneer of the technique named cortico-cortical evoked potentials (CCEPs) some 30 years later.25 These anatomical and physiological advances eventually convinced Talairach that he should add a significant section on SEEG and neurophysiology to his stereotactic anatomical atlas of the telencephalon in 1967.6

10   | SECTION I. INTRODUCTION

The Institut National de la Santé et de la Recherche Médicale Research Unit entitled “Stereotaxic Functional Exploration and Surgical Treatment of the Epilepsies” headed by Talairach was created in 1970. This new facility helped Bancaud to approach some basic problems in the electrophysiological investigation of the human epilepsies and set new trends in the characterization of frontal lobe seizures. Bancaud and Talairach paid more attention and invested much more time to write books than journal articles. They considered that the challenge posed by human intracerebral recordings, the new concepts on epilepsies they were generating, and their consequences for surgical strategy were so complex that they could not be thoroughly described and properly discussed in a limited article format. They had themselves been inspired by the detailed clinical experience handed on from their predecessors. As a matter of proof, at the end of my first interview with Bancaud in 1971, I was given the five books of Penfield to read before getting back to him. In the same period, Bancaud returned to the questions of scalp EEG interpretation raised in his doctoral thesis and clarified many of them on the basis of data accumulated from comparisons between EEG and SEEG. The 1973 book EEG et SEEG dans les Tumeurs Cérébrales et l’Épilepsie was especially dedicated to electroencephalographers and collected unique and important data on simultaneous recordings of EEG and SEEG during numerous investigations of cerebral tumors and epilepsies in the stereotaxy “chapel.” Nowadays, it remains a valuable and useful guide to understanding what we record and what we miss in surface recordings of slow waves, spikes, and seizures.26 Between 1970 and 1973, with the help of Stéphane Geier, the recording period of SEEG lengthened, and from “acute,” SEEG became “chronic.” Technically speaking, intracerebral electrodes were significantly reduced in size and adapted to the new mode of recording while the patient was freely moving. Telemetry was used, as well as video-recording with EEG and video signals being retrospectively matched.27 These new types of recording out of the operating room allowed a much better study of gestural automatic behavior that characterize certain complex partial seizures. In particular gestural manifestations, correlated with anterior frontal lobe paroxysmal discharges, were identified.28 Their phenomenology, duration, and mode of association differentiated them from gestural automatisms of temporal lobe origin.29 Some observations made at this time suggested an epileptogenic zone distributed between frontal and temporal lobe areas, leading to the concept of frontotemporal epilepsy, and to a concomitant change in surgical strategy. In the search for mechanisms of “automatisms,” Bancaud and his coworkers reported the behavioral effects of anterior cingulate gyrus stimulation which essentially consisted of complex hand–mouth coordinated movements with mood alteration toward disinhibition.9,30 The continuing collaborative work of Jean Talairach and Jean Bancaud culminated in a global report presented in 1974 to the Société de Neurochirurgie de Langue Française, entitled “Approche Nouvelle de la Neurochirurgie de I’Epilepsie,” and published as a supplement volume of the Neurochirurgie journal.31 This presentation specified the successive and multidisciplinary steps of the method and made clear the importance of making the surgical operation fit the seizure semiology through a meticulous and rational correlative approach. So remarkably extensive was the range of research and curiosity of Talairach and Bancaud that it is difficult to highlight the major elements of their scientific legacy. Surely one of their most important contributions is the emphasis they have placed on the value of clinical semiology and its integration through the “anatomo-electro-clinical correlations” of a patient’s seizures. The body of work that has been produced from these principles has transformed what was at the beginning a technique of stereotactic functional exploration into a general method of studying a patient with epilepsy. In a sense, they have fashioned a tool that has turned out to be fundamental, as proven by the fact that it has survived the development of imaging and other noninvasive techniques.32 Indeed, not only has it survived, but it has framed a preinvasive rationale for the current imaging techniques by providing a referential system perfectly suited to the investigation of this anatomo-functional crossroad.33

1. The History and Principles of Stereo EEG   | 11 

SUMMARY A serendipitous encounter between Jean Talairach and Jean Bancaud in the late 50s marked the origins of a new conception in presurgical exploration of epilepsy. They named it stereoelectroencephalography or SEEG, meaning that it was based on a multidimensional (supplied by stereotaxy) recording of multiple brain areas (through multilead depth electrodes) simultaneously. This novel method allowed for separating a presurgical localization from a surgical treatment phase. A closed skull setup enabled intracerebral seizure recording for the first time in a fully interacting patient. SEEG interpretation was based on ictal anatomo-electro-clinical correlations. A better comprehension of seizure build-up in the human brain led to major pathophysiological advances. SEEG-guided epilepsy surgery was performed only in Paris during the first 25 years, and then adopted in Switzerland and Canada. Ten years later, its practice developed in France (Rennes and Grenoble), and also in Italy (Milan). Its recent worldwide expansion confirms its universally recognized rationality and its perfect adjustment to brain imaging.

KEY REFERENCES Only

key references appear in the print edition.

The

full reference list appears in the digital product

found on http://connect.springerpub.com/content/book/978-0-8261-3693-0/part/part01/chapter/ch01

  1. Talairach J. Souvenirs Des Etudes Stereotaxiques Du Cerveau Humain: Une Vie, Une Equipe, Une Methodologie: L’ecole de Sainte-Anne. John Libbey Eurotext; 2007. 15. Talairach J, Tournoux P. Co-Planar Stereotaxic Atlas of the Human Brain. 3-Dimensional Proportional System: An Approach to Cerebral Imaging. Georg Thieme Verlag; 1988. 19. Bancaud J, Talairach J, Bonis A, et al. La Stéréo-Electroencéphalographie dans l’épilepsie. Informations neurophysiopathologiques apportées par l’investigation fonctionnelle stéréotaxique. Masson & Co; 1965. 29. Geier S, Bancaud J, Talairach J, et al. The seizures of frontal lobe epilepsy. A study of clinical manifestations. Neurology. 1977;27(10):951–958. doi:10.1212/wnl.27.10.951 30. Talairach J, Bancaud J, Geier S, et al. The cingulate gyrus and human behaviour. Electroencephalogr Clin Neurophysiol. 1973;34(1):45–52. doi:10.1016/0013-4694(73)90149-1 31. Talairach J, Bancaud J, Szikla G, et al. [New approach to the neurosurgery of epilepsy. Stereotaxic methodology and therapeutic results. 1. Introduction and history]. Neurochirurgie. 1974;20 (suppl 1):1–240. 33. Chauvel P, Gonzalez-Martinez J, Bulacio J. Presurgical intracranial investigations in epilepsy surgery. In: Levin KH, Chauvel P, eds. Handbook of Clinical Neurology. Vol 161. Elsevier; 2019:45–71. doi:10.1016/B978-0-444-64142-7.00040-0 34. Bancaud J, Brunet-Bourgin F, Chauvel P, Halgren E. Anatomical origin of déjà vu and vivid “memories” in human temporal lobe epilepsy. Brain J Neurol. 1994;117(Pt 1):71–90. doi:10.1093/ brain/117.1.71

Section II.  Noninvasive Evaluation

2 Phase I Evaluation

Allyson A. Pickard and Christopher T. Skidmore

KEY CONCEPTS • The epileptogenic zone (EZ) is the region of brain that causes epilepsy and, once removed, can provide a cure. • The EZ cannot be evaluated directly, but rather approximated through evaluation of alternate regions for which testing is available. • These regions include the symptomatogenic zone, the irritative zone, the ictal onset zone, the functional deficit zone, the epileptogenic lesion, and the eloquent cortex. • Sometimes the data obtained from study of these zones are discordant and a more invasive evaluation is required.

INTRODUCTION Epilepsy surgery is a powerful treatment that can offer a cure for patients suffering from seizures. The main goal of a presurgical workup is to identify who may benefit from surgery. This is done with precise seizure localization and determination of any risks or deficits from the proposed surgery. The first step to this goal is through noninvasive testing. Several so-called "zones of the brain"—to distinguish the term from anatomical regions—and their relationship to one another are to be defined and used to help guide an evaluation for surgery. They include the symptomatogenic zone, the irritative zone, the ictal onset zone, the functional deficit zone, the epileptogenic lesion, and the eloquent cortex.1 The final region, called the "epileptogenic zone" (EZ), is the ultimate goal and is defined as the minimal area of brain to be removed or disconnected to render the patient seizure free. The EZ cannot be directly tested and is therefore theoretical in nature, approximated by careful analysis of the other areas for which testing modalities are available. In this chapter we illustrate how an understanding of these zones helps to approximate the EZ and describe the common methodologies utilized to assess each zone. This chapter is not meant to be an exhaustive discussion of each area and several of the techniques/technologies that will be mentioned will be discussed in more detail in the following chapters. To help give context and understanding to these regions throughout this chapter, consider a straightforward, hypothetical patient:

16   | SECTION II. NONINVASIVE EVALUATION

A 46-year-old right-handed man presents with medically refractory epilepsy since onset at age 28. He has no risk factors for epilepsy. He continues to have seizures on high doses of zonisamide and oxcarbazepine. He has previously taken levetiracetam, valproate, and lacosamide at their maximal doses without improvement. He has focal impaired awareness seizures lasting about 1 minute once per month and has focal to bilateral tonic-clonic seizures about once per year. His workup will highlight the important points of a noninvasive presurgical evaluation.

SYMPTOMATOGENIC ZONE The symptomatogenic zone is defined as the region which generates the initial seizure semiology. Note that this is often not the same region as an ictal onset as a seizure must frequently evolve and spread after onset for a clinical change to occur. However, overlap between the two regions is possible. This zone is best identified by obtaining a detailed description of the behavior experienced by the patient or seen by others. This can be achieved by both a careful and sequential history of the seizure as well as a video of the seizure itself, whether captured by a family witness or with dedicated video-monitoring equipment in an epilepsy monitoring unit. Attention to key components can help structure analysis. The presence of an aura, whether it be somatosensory, visual, auditory, and/or focal clonic activity at onset, is helpful as it is the first symptom of a seizure and occurs before a seizure has significantly spread.2 Each of these possible symptoms implicates a very specific region of the cortex as the primary area of concern, though there are potential alternative regions as well. Somatosensory auras, for instance, would suggest a possible primary sensory cortex localization or perhaps the second sensory cortex located in the posterior regions of the superior temporal gyrus. Therefore, it may give insight into a more precise symptomatogenic zone than a seizure that occurs without an aura. There are numerous lateralizing and localizing signs that, when identified, can also help determine the region of interest or at least lateralize the location of the seizure onset. A detailed description of these signs is beyond the scope of this chapter, but a very nice summary of the most common signs can be found in a review article written by Dr. Loddenkemper and Dr. Kotagal.3 Common semiologic features associated with various cortical regions of onset have been summarized in Table 2.1 and will be discussed in what follows.4 Finally, the early impairment of language without alteration of awareness or consciousness helps to confirm a seizure’s initial presence in the dominant hemisphere, an important consideration when planning an invasive EEG implant and eventually determining the extent of a surgical operation.

Temporal Lobe Seizures The temporal lobe is the most common location for seizures to begin. The mesial temporal lobe frequently has a semiology that begins with a viscerosensory or fear aura. Patients may describe a sudden sense of dread or a feeling of déjà vu. Oral and manual automatisms are common though secondary generalization is rare and often occurs late in the seizures if present. The lateral temporal lobe, by contrast, has frequent rapid secondary generalization. Additionally, an auditory aura is typically described and due to its proximity to Wernicke’s area, the patient may become aphasic, which typically becomes apparent postictally.

Frontal Lobe Seizures Seizures arising from the frontal lobes can be varied, with several locations each resulting in a different semiology.5 A supplementary sensorimotor seizure for instance presents with focal asymmetric tonic posturing, head version typically contralateral to the symptomatogenic

2. Phase I Evaluation   | 17 

TABLE 2.1  LOCALIZING SEMIOLOGY CEREBRAL LOBE

SEMIOLOGY

  Primary motor

Focal motor symptoms

  Supplementary motor

Focal asymmetric tonic posturing

 Orbitofrontal

Olfactory hallucinations Complex motor automatisms

 Opercular

Salivation Swallowing Gustatory hallucinations

 Cingulate

Mood changes Vocalization Similar findings to any other area

Frontal

Temporal  Mesial

Viscerosensory or fearful auras Oral and manual automatisms Rare secondary generalization

 Lateral

Auditory auras Aphasic seizures (dominant hemisphere) Rapid secondary generalization

Insular

Throat or neck tightening Painful somatosensory auras

Parietal

Somatosensory auras

Occipital

Visual auras Blinking

zone, speech arrest, and vocalization. An orbitofrontal seizure, by contrast, may present with complex motor automatisms and olfactory hallucinations. Occasionally, these automatisms are dramatic and atypical, leading to an erroneous diagnosis of psychogenic nonepileptic seizures. Care must be taken to appreciate that the semiology is stereotyped in these patients.

Parietal Lobe Seizures Due to the proximity to the primary and supplementary sensory cortices, parietal lobe seizures frequently begin with a somatosensory aura such as tingling or numbness. These are typically contralateral to the symptomatogenic zone, though they may be bilateral. Very frequently, there is spread to the adjacent motor cortex resulting in motor involvement as well. It should be noted, however, that a painful somatosensory aura is more typical of an insular lobe seizure and its presence should lead to consideration of this alternate region.

Occipital Lobe Seizures The occipital lobe controls and processes visual information and as such seizures from this lobe often have visual auras. Other symptoms could include ictal blindness, ocular movement, and blinking. There is also frequently rapid spread to other lobes, resulting in respective contributions to semiology that can cause ambiguity and misidentification for the initial seizure semiology.

18   | SECTION II. NONINVASIVE EVALUATION

Returning to the hypothetical patient, the semiological description obtained after three typical seizures were captured on video is as follows: He sits in bed working on a computer when he suddenly stops and looks up. He presses the seizure button and states he is about to have one and states he is having a déjà vu sensation. He begins to make chewing movements and rubs the fingers of his left hand. The nurse enters the room and the patient initially is unable to speak but does follow simple motor commands. After approximately 30 seconds, he stops following all commands. He rubs his thigh. A code word is given. After approximately 1 minute, the patient again interacts with the nurse. He does not remember the code word nor the chewing or rubbing movements. This is a typical semiology of a mesial temporal lobe seizure with an aura of déjà vu and the presence of automatisms. The early retained awareness with loss of speech is consistent with a seizure in the hemisphere dominant for language which for the right-handed patient is the left side of the brain. In summary, the patient’s symptomatogenic zone is the left mesial temporal lobe. We could summarize the semiology as: psychic (déjà vu) aura → aphasic seizure → automotor seizure, capturing the most pertinent aspects of the semiology.

IRRITATIVE ZONE The irritative zone is the cortical region that generates interictal epileptiform activity. This is the location of the spikes and sharp waves detected on an EEG or magnetoencephalography (MEG). Not every patient with epilepsy generates this abnormality and occasionally they can be quite rare, necessitating prolonged EEG monitoring for detection. Epileptiform discharges are more commonly seen during nonREM sleep than during wakefulness and REM sleep. However, epileptiform discharges from REM sleep and wakefulness may be more geographically restricted and have a potentially higher degree of correlation with the ictal onset zone.6 Discharges may also become more prevalent after seizures. Precise localization of the epileptiform discharges, and therefore the irritative zone, is often limited by the technical limitations of the EEG. To detect a discharge on scalp EEG, approximately 6 cm2 of cortex must be involved with discharges of a smaller area not appearing at all on scalp EEG tracings.7 Furthermore, each discharge must pass through the meninges and bone of the skull, which may distort the localization of the activity. A discharge may be seen more prominently in the anterior temporal region while another discharge may be maximal in the midtemporal region within the same study.8 Other modalities may be required to help reconcile if these two locations are in fact generated by one focus, or if two nearby foci must be considered. A somewhat more precise map may be generated with a dense array EEG which can allow for up to 256 channels modelling the suspected source on the patient’s own MRI. Electrical source imaging can then be utilized to account for the impacts the bone and soft tissues have on the electrical potential to suggest a probable cortical source. The hypothetical patient’s interictal EEG is demonstrated in Figure 2.1. In the first image a sharp wave is seen with maximal negativity at T1. In the second image, recorded a few hours later, the maximal negativity is seen at T3. Note that in both images the field is broad, extending to the temporal-occipital region. The patient’s irritative zone is best described as being located in the left anterior to midtemporal region. An alternative to scalp EEG in determining the irritative zone is MEG. An epileptic discharge generates an electrical dipole that is perpendicular to the cortical surface. This dipole, in turn, generates a magnetic field that can be detected and localized on a brain MRI.9 Unlike scalp EEG, MEG is not subject to the distortions created by passing through meninges and skull. It does, however, require somewhat frequent epileptic discharges to reliably localize the irritative zone. Unfortunately, this technology is limited by availability and decreased recording time as compared to continuous EEG recordings.

2. Phase I Evaluation   | 19 

FIGURE 2.1  Interictal EEG findings of the hypothetical patient. There is variability of the maximal

negativity between left anterior (T1) and mid (T3) temporal.

ICTAL ONSET ZONE The same modalities used to identify the irritative zone are used to help characterize the ictal onset zone, or the area of brain where a seizure first begins electrically. The same limitations present for EEG for evaluation of the irritative zone again limit the evaluation of ictal onset. Focal aware seizures may not appear on scalp EEG due to a small area of cortical involvement. Frontal seizures may poorly lateralize or have only subtle ictal changes. Seizures with a prolonged aura may have a delayed EEG onset, reflecting a spread pattern rather than the onset zone of interest. It is important to remember that even a precise localization of ictal onset does not guarantee seizure freedom as the EZ may go beyond the recorded seizure onset. This can be a reflection of the limitations of EEG, the multifactorial nature of the EZ, or both. Many epilepsy centers use an additional modality to help confirm the ictal onset zone using ictal single-photon emission computed tomography (SPECT).10 This technique uses a radiotracer to help measure cerebral blood flow, which is increased in an area with high metabolism, such as the cortex involved in seizure onset. Ideally, the maximum blood flow and metabolism is seen at the origin of a seizure and the areas where a seizure initially spreads. This requires injection of the isotope within seconds of seizure onset. A delay in injection or a rapid ictal spread pattern will significantly lower the accuracy of the study. Ictal SPECT does, however, offer the advantage over EEG of evaluating all areas of the brain including areas often missed by EEG such as deep cortical structures. Due to the limited recording time inherent to MEG, it is rare to record seizures during a MEG study. However, seizures that are recorded can be localized utilizing the same principles for localizing interictal epileptiform discharges. An ictal EEG of the hypothetical patient is shown in Figure 2.2. Note that the patient felt his first symptoms several seconds before an ictal change is seen on EEG. The final report gives the following electrographic description: A 6–8 Hz rhythmic discharge is seen maximal at T1 for 9 seconds. This evolves to slower delta rhythms (3–4 Hz) with spread to the left posterior temporal and parietal regions.

20   | SECTION II. NONINVASIVE EVALUATION

FIGURE 2.2  Ictal EEG findings of the hypothetical patient. The first ictal changes are seen maximally

in the left anterior temporal region before evolution across the left temporal region and hemisphere occurs.

There is further evolution to 2 Hz rhythmic spiking activity before an abrupt offset is seen. The electrographic seizure lasts 52 seconds.

EPILEPTOGENIC LESION A structural abnormality that is thought to give rise to seizures is known as the epileptogenic lesion. Nearly every patient receives some form of head imaging after a first time seizure, initially using CT. This will help identify more extensive lesions, such as high-grade tumors as well as hemorrhages, that may have caused the seizure. Patients with longstanding epilepsy, however, require a more detailed image to identify more subtle lesions, such as mesial temporal sclerosis or focal cortical dysplasia. An MRI is most commonly used providing higher resolution, correlating with stronger magnetic fields, for the best quality image. As the imaging technology improves, smaller abnormalities are able to be identified including heterotopias, encephaloceles, and other abnormalities of cerebral development. PET utilizing specific ligands have also been developed to identify functional or specific structural lesions such as cortical tubers in individuals with tuberous sclerosis. This will be covered in the chapter on PET imaging. Figure 2.3 shows the coronal MRI of the hypothetical patient. Left mesial temporal sclerosis is demonstrated by the hyperintensity and atrophy of the hippocampus.

FUNCTIONAL DEFICIT ZONE A patient with medically refractory epilepsy may present with a chronic clinical deficit that persists in the interictal period, such as a poor category-specific memory or hemiparesis. The area responsible for this interictal finding is the functional deficit zone. This may be due to a visible structural lesion, abnormal neuronal networks, or inadequate neuronal signaling in the

2. Phase I Evaluation   | 21 

FIGURE 2.3  Coronal T1-weighted (A) and T2-weighted (B) fluid-attenuated inversion recovery MRI

of the hypothetical patient showing severe left mesial temporal sclerosis.

region. It may become larger or more significant over time as seizures continue. Neurological examination may be able to detect more severe deficits such as sensory loss or weakness. Deficits that may impact the patient’s ability to perform their job, such as moderate memory loss, may not be detected on routine examination though the patient may report declining work performance. A detailed neuropsychological assessment may be used to identify and describe these more nuanced deficits. A better understanding of any functional deficits can be used to both predict an area of brain that may be affected as well as any impact of surgery. General intelligence including verbal, perceptual, and general intelligence quotient modalities may be tested. Verbal and visual memory testing can be helpful to identify specific memory deficits while testing naming and verbal fluency may uncover a language abnormality. Screening for anxiety and depression are also frequently done as these may significantly impact patient quality of life during the presurgical evaluation and postsurgical recovery if unrecognized and untreated.11 The detailed neuropsychological report of the hypothetical patient includes assessments of his general intelligence, visual memory, verbal memory, naming, verbal fluency as well as anxiety and depression indexes. An excerpt from the report follows. A general ability index (GAI) is used as a substitute for a Full Scale Intelligence Quotient and is average at 102 (55th percentile). His verbal memory using a list delayed recall and a prose passage immediate recall for assessment was borderline impaired in the 7th percentile. His visual memory was average in the 56th percentile after testing immediate and delayed recall of object designs. His naming was found to be impaired at the 4th percentile when he was only able to name 48 out of 60 objects. Verbal fluency was similarly impaired with phonemic fluency at the 10th percentile and semantic fluency at the 1st percentile. Anxiety and depression were evaluated using standardized assessments and found to be minimal with the patient reporting mild symptoms. The impairment of naming and verbal fluency suggests dysfunction in the dominant hemisphere. This is further narrowed with the observed impairment of verbal memory, suggesting dominant temporal lobe dysfunction. Like the ictal SPECT showing hypermetabolism during a seizure, the interictal PET can show hypometabolism between seizures in the same area. Even if a structural abnormality is not seen on a MRI, a hypometabolic area on an interictal PET scan indicates an area of

22   | SECTION II. NONINVASIVE EVALUATION

FIGURE 2.4  PET imaging of the hypothetical patient. This coronal image shows decreased radio-

tracer uptake in the left temporal lobe, maximally medially, indicated by the arrows. This corresponds to hypometabolism in this region.

dysfunction. The PET scan indicating left temporal lobe hypometabolism found for the hypothetical patient is shown in Figure 2.4. Interictal SPECT may also be used; however it has lower resolution than PET leaving its utility in establishing a baseline image for an ictal SPECT study limited. Some researchers have found the fluorodeoxyglucose-PET imaging with concordant semiology, interictal and ictal EEG for mesial temporal onset epilepsy is as positive a predictor of surgical success as mesial temporal sclerosis on MRI imaging.12

ELOQUENT CORTEX As important as the accurate estimation of the EZ in the presurgical evaluation is for seizure freedom, the identification of eloquent cortex is equally as important to determine the functional outcome of surgery. A patient who loses the ability to understand or speak as a result of epilepsy surgery may become seizure-free but may remain disabled and unable to return to work. As such, part of the presurgical evaluation includes identification of the eloquent cortex to help understand and define boundaries for surgery. Functional MRI (fMRI) can be used to identify eloquent cortex through evaluation of neuronal networks and any restructuring of these networks. Similar to the SPECT and PET modalities, this study evaluates the blood flow and metabolism of the brain. With fMRI, however, these changes are measured in response to certain tasks, such as motor function, movement, or language.13 A patient is asked to perform a movement, or remember a set of words, and the resultant change in cerebral blood flow found on the image indicates the area of brain responsible for that task. After a series of tasks, a map of functional cortex can be generated. Any deviation from the expected norms can be used in the interpretation and evaluation of the functional deficit zone previously discussed. These images can be combined with MRI diffusion tensor imaging images, which permit the visualization of structural connectivity in the form of white matter tracks traveling between cortical regions of interest. This helps to identify white matter bundles which need to be avoided during surgery. Figure 2.5 shows a normal fMRI with intact language areas identified in the left hemisphere. For the hypothetical patient, any surgical plans will need to consider preservation of these areas. Evoked potentials and transcranial magnetic stimulation are additional noninvasive studies that could be performed. Stimulation is applied and the resultant action or signal can be used

2. Phase I Evaluation   | 23 

FIGURE 2.5  fMRI of the hypothetical patient. The MRI sequences have been superimposed with

evidence of cerebral activation during a sentence completion task, measured as changes in blood flow. The bottom right image is a reconstruction showing all language activation. fMRI, functional MRI; L, left; R, right.

to map eloquent cortex. The Wada test (named after Juhn Wada, also known as "intracarotid ambobarbital" test) is an invasive test often performed prior to the implantation of intracranial electrodes and for that reason is mentioned here. It involves the injection of an anesthetic agent such as amobarbital into the carotid artery to selectively anesthetize the cerebral hemisphere ipsilateral to the injection. This permits the neurologist or neuropsychologist to test the remaining “awake” hemisphere for memory, language, and motor function.

EPILEPTOGENIC ZONE AND SURGICAL PLANNING With evaluation of the previously discussed zones, summarized in Table 2.2, the theoretical EZ can be constructed. A minimum workup for every patient to provide an adequate depiction of this zone includes a history and physical examination, video-EEG monitoring, MRI, and neuropsychological testing often referred to as a "Phase 1 evaluation." Many centers find additional testing such as the PET, MEG, or fRMI to help further refine the hypothesis of the EZ. Often the patient’s individual findings are summarized and discussed at a surgical conference with epileptologists, neurosurgeons, and neuropsychologists present. Table 2.3 shows the findings of the hypothetical patient. His findings strongly suggest an EZ in the left anterior temporal region. Some may argue that for this patient, further workup may not be needed

24   | SECTION II. NONINVASIVE EVALUATION

TABLE 2.2  FACTORS IN EVALUATION OF THE EPILEPTOGENIC ZONE ZONE EVALUATED

DEFINITION

MODALITIES USED

Irritative zone

Region generating interictal epileptiform activity

EEG MEG

Ictal onset zone

Region where a seizure first begins electrically

EEG Ictal SPECT

Epileptogenic lesion

Structural abnormality thought to MRI brain give rise to seizures CT head PET

Functional deficit zone

Chronic clinical deficit that persists interictally

Neuropsychological testing PET Interictal SPECT

Eloquent cortex

Functional region of the brain, such as for motor or language

fMRI MRI DTI Evoked potentials Wada

EZ

Region of brain causing epilepsy

All of the above

Symptomatogenic zone

Region generating initial ictal semiology

Clinical history Video review

DTI, diffusion tensor imaging; EZ, epileptogenic zone; fMRI, functional MRI; MEG, magnetoencephalography; SPECT, single photon emission computerized tomography.

TABLE 2.3  HYPOTHETICAL PATIENT EVALUATED ZONE

MODALITY

LOCALIZATION

Irritative zone

EEG

Left anterior to midtemporal lobe

Ictal onset

EEG

Left anterior temporal lobe

Epileptogenic lesion

MRI

Left mesial temporal lobe

Functional deficit

PET

Left anterior temporal lobe

Neuropsych

Left temporal lobe

fMRI

Left hemisphere language dominance

Symptomatogenic zone

Eloquent cortex EZ hypothesis

History/Video

Left mesial temporal lobe

Left anterior/mesial temporal region

EZ, epileptogenic zone; fMRI, functional MRI. Note: The data obtained from the hypothetical patient’s workup is concordant and gives rise to a single leading hypothesis of the EZ.

and that he could proceed to resective surgery with an anterior temporal lobectomy.12 The extent of his resection, however, may need to be limited by the posterior Wernicke’s area to preserve language. Additionally, due to his verbal memory impairment, he is at risk of further verbal memory decline with an anterior temporal lobectomy.14 This may lead some centers to consider laser interstitial thermal therapy (or laser ablation) to target just his mesial temporal sclerosis in an attempt to preserve memory outcome and speed patient recovery following the procedure. Alternately, some centers may wish to more precisely evaluate the EZ using intracranial EEG to help minimize the extent of resection. Not every patient is as straightforward. Rarely is there complete agreement between the identified zones. Sometimes two or more hypotheses are generated from the noninvasive

2. Phase I Evaluation   | 25 

TABLE 2.4  ALTERNATE PATIENT EVALUATED ZONE

MODALITY

LOCALIZATION

Irritative zone

EEG

Bilateral independent anterior temporal lobes

Ictal onset

EEG

Left mid to posterior temporal lobe

Epileptogenic lesion

MRI

Right frontal operculum

Functional deficit

PET

Right frontal and temporal lobe

Neuropsych

Left temporal lobe

fMRI

Left hemisphere language dominance

Symptomatic zone

Eloquent cortex

History/Video

EZ hypotheses

Mesial temporal lobe

Left anterior/mesial temporal region Right anterior/mesial temporal region Right frontal opercular/insular region

EZ, epileptogenic zone; fMRI, functional MRI. Note: The data obtained from the alternate patient is discordant, leading to several differing hypotheses of the EZ.

evaluation due to discordant data. Consider the findings of an alternate patient in Table 2.4. With this evaluation, the MRI showed a right frontal abnormality suggesting a potential epileptogenic lesion in this area. However the symptomatogenic zone determined by clinical history and video analysis supports a mesial temporal onset. There is additional supporting evidence for each of these potential locations as well as other possibilities not completely assessed with the studies performed, such as the bilateral irritative zone or the left mid to posterior temporal ictal onset zone findings. Which hypothesis is correct? To help answer this question and reconcile the findings, an invasive study using intracranial EEG is needed in the presurgical evaluation.

SUMMARY The EZ, if correctly hypothesized, can allow surgery to provide a cure for epilepsy. Approximation of this ultimate goal is done through analysis and understanding of several alternate regions including the symptomatogenic zone, the irritative zone, the ictal onset zone, the functional deficit zone, the epileptogenic lesion, and the eloquent cortex. There are many noninvasive tools for evaluation of the alternate zones including history, video, imaging, EEG, and neuropsychological assessment. Occasionally these results, when examined, provide a clear picture of the likely EZ. However, they are often not in complete agreement and a more invasive approach is required to provide the best hypothesis.

KEY REFERENCES Only

key references appear in the print edition.

The

full reference list appears in the digital product

found on http://connect.springerpub.com/content/book/978-0-8261-3693-0/part/part02/chapter/ch02

1. Lüders H, Comair YG. Epilepsy Surgery. 2nd ed. Lippincott Williams & Wilkins; 2001. 2. Lüders H, Fernandez-Baca Vaca G, Akamatsu N, et al. Classification of paroxysmal events and the four-dimensional epilepsy classification system. Epileptic Disord. 2019;21(1):1–29 doi:10.1684/ epd.2019.1033 4. Skidmore CT. Adult focal epilepsies. Continuum (Minneap Minn). 2016;22(1 Epilepsy):94–115. doi:10.1212/CON.0000000000000290

26   | SECTION II. NONINVASIVE EVALUATION 12. LoPinto-Khoury C, Sperling MR, Skidmore C, et al. Surgical outcome in PET-positive, MRI-negative patients with temporal lobe epilepsy. Epilepsia. 2012;53(2):342–348. doi:10.1111/j.15281167.2011.03359.x 14. Stroup E, Langfitt J, Berg M, et al. Predicting verbal memory decline following anterior temporal lobectomy (ATL). Neurology. 2003;60(8):1266–1273. doi:10.1212/01.wnl.0000058765.33878.0d

3 Advanced MRI Imaging Eliane Kobayashi

KEY CONCEPTS • MRI is a core part of presurgical investigation, but it needs to include sequences that maximize lesion detection to improve identification of targets for intracranial EEG (ICEEG) electrodes placement or resection. • A gold-standard protocol for MRI acquisition in epilepsy (named "Harmonized Neuroimaging of Epilepsy Structural Sequences" [HARNESS]-MRI) was recently proposed by the Neuroimaging Taskforce International League Against Epilepsy (ILAE) Commission. • Visual analysis by trained personnel might be sufficient to detect structural abnormalities, but postprocessing and computational techniques might further increase the yield of lesion detection. • MRI morphometry, including volume-based and surface-based analyses, should be performed according to clinical suspicion and interpreted with basis of clinical information. • With appropriate imaging, most MRI-negative patients could be considered cryptogenic rather nonlesional, as imaging technologies evolve both in acquisition, analysis, and classification. • There is a growing role for imaging classifiers and multimodal MRI profiling based on features extracted from different MRI sequences.

INTRODUCTION The advent of MRI in epilepsy is one of the most impactful in neurology. It added to the dichotomy of focal versus generalized epilepsy, the classification of lesional and nonlesional (or cryptogenic) epilepsy, bringing crucial information for treatment decision in patients who suffer from drug-resistant epilepsy (DRE). Finding an underlying structural lesion shifts an electroclinical diagnosis toward potentially surgically treatable etiologies and thus, remediable seizures, which one would aim to identify early and to treat precisely. In clinical practice, patients with newly diagnosed epilepsies (and not only DRE patients in presurgical workup) should also undergo a proper MRI epilepsy protocol, as this can often clarify epilepsy types and guide treatment decisions timely. In presurgical investigation of DRE, ideally, we would like to identify a well-defined and circumscribed structural lesion, in a noneloquent cortical area and well correlated with the electroclinical information. Ultimately, that lesion could be completely and safely resected, later to have a confirmed diagnosis with ex vivo neuropathology.

28   | SECTION II. NONINVASIVE EVALUATION

Despite the multiple subtypes of MRI images available nowadays, this chapter, in the context of this book, focuses only on anatomical (structural) MRI for identification of lesions. This is just one contribution, in the broader understanding of epilepsy diagnosis and its natural evolution, resulting from MRI applications. It is not intended to dismiss valuable information provided by other MRI sequences that can be added to the standard protocol and that contribute to the understanding of lesion progression and network abnormalities.

AN OPTIMIZED MRI EPILEPSY PROTOCOL To optimally increase the yield of MRI investigation in surgical candidates, proper imaging protocols need to be readily available in MRI scanners worldwide and personnel reporting/ evaluating these images need to have proper training and knowledge to look for meaningful findings (i.e., epileptogenic lesions). Under that perspective, the epilepsy team holds in its hands detailed information derived from extensive investigation including video-EEG monitoring and neuropsychology evaluation, positioning itself at the best place to unravel lesions from MRI. Thus, the admission to the epilepsy-monitoring unit is a crucial opportunity to gear towards further testing, including repeat MRI scanning and postprocessing MRI analyses, to find lesions otherwise not previously reported or found in earlier scans. The benefit of repeating MRI scanning with the appropriate protocol surpasses any additional cost it brings to the healthcare system, bringing faster and more efficient diagnosis clarification and surgical decisions, both by using higher field magnets than previous investigations (i.e., 1.5T vs. 3T, and in some scattered centers also 7T MRI) and by acquiring better imaging sequences including use of different head coils that allow fast parallel imaging. These protocols are most often readily available at scanners in tertiary academic epilepsy centers guiding epilepsy surgery and ICEEG, but it might be underutilized elsewhere. The Neuroimaging Taskforce of the ILAE Diagnostic Methods Commission recently published recommendations aiming to assist physicians in determining minimum MRI requirements for the investigation of epilepsy patients.1 Moreover, the task force reviewed currently available postprocessing imaging techniques that could optimize lesion detection, making recommendations on how to apply them and which patients to further investigate. In this guideline, which took in consideration the importance of availability of different scanners worldwide, emphasis was put on three-dimensional (3D) acquisitions using isotropic voxels that allow multiplanar reconstruction. This proposed ILAE gold-standard protocol for epilepsy can be applied both to 1.5T and 3T MRI scanners, as well as to adult and pediatric patients, and it has been named the HARNESS-MRI protocol.1 Compared to previous ILAE recommendations, this new MRI guideline eliminates the need for specification of an MRI protocol for temporal lobe epilepsy (TLE) versus extra-TLE patients, as most images are 3D acquisitions that could be later reconstructed in any desired angulation to serve both patient groups. The two-dimensional–coronal-thin slices focused on the temporal lobes and oriented perpendicularly to the hippocampal long axis are still used, being very valuable in busy clinical settings and bedside evaluations. Similar to previous recommendations, anatomy and sulcal/gyral morphology should be assessed on a 3D T1-weighted gradient echo MRI sequence (isotropic 1 mm voxels with no interslice gap is the standard main parameter), such as magnetization-prepared rapid gradient echo, 3D spoiled gradient echo, and 3D turbo field echo protocols. 3D fluid-attenuated inversion recovery (FLAIR) is the choice of images for evaluating signal abnormalities, in particular hyperintense transmantle signs of focal cortical dysplasias (FCDs; Figure 3.1), but also those of hippocampal sclerosis (HS). Acquisition time in general can be optimized with parallel imaging and this is very important considering patients’ capacity to stay still for long periods of time.

3. Advanced MRI Imaging   | 29 

FIGURE 3.1  3D isotropic T1-weighted and fluid-attenuated inversion recovery images from 3T MRI

scanner reveals a right frontal focal cortical dysplasia at the bottom of the sulcus with a transmantle sign. R, right.

WHAT MRI “CAN SEE” A well-defined and characterized lesion can provide good enough evidence for the location and extent of the epileptogenic zone to be resected, thus avoiding the need for ICEEG investigation in some patients. There is still a debate as to whether, in the presence of such clearly identified lesions, patients should still undergo ICEEG, but when an extensive or multifocal abnormality is found, this might be crucial prior to surgical decisions. MRI might reveal only the “tip of the iceberg,” and the lesion could be much more extensive than the structural abnormality readily detectable. That is the reason why acquisition of multiple MRI sequences and analyzing them in combination—multimodal imaging profiling—could increase the yield for detection and estimate of lesion extension. In contrast, a “nonlesional” MRI should not rule out an underlying abnormality or preclude surgical decisions, in particular for extra-TLEs, in which malformations of cortical development such as FCDs can be invisible to qualitative (i.e., visual) MRI analysis and in some cases, even to postprocessing techniques. This is particularly true for neuropathology-confirmed FCD IIa.3,4 It is therefore important to apply and take advantage of imaging processing and computational approaches that can unravel subtle abnormalities which could potentially be targeted by surgery.5–7 Whereas finding an MRI abnormality can be confirmatory in the context of clinical investigation, the high yield of MRI investigations requires critical evaluation and classification by well-trained personnel looking at appropriate imaging protocols, as they can reveal both incidental and potentially epileptogenic lesions.8 The presence of an MRI-identifiable lesion can fast-forward such a process, but lesions do not always unravel as a well-circumscribed structural abnormality but rather as extensive or multifocal structural lesions such as polymicrogyria, hemimegalencephaly, and tuberous sclerosis (Figure 3.2).

MRI FOR IDENTIFYING LESIONS IN MESIAL TEMPORAL LOBE EPILEPSY In mesial temporal lobe epilepsy (MTLE), structural neuroimaging investigation aims to identify MRI evidence of mesial temporal sclerosis, of which HS is the major feature. There are

30   | SECTION II. NONINVASIVE EVALUATION

FIGURE 3.2  Right hemimegalencephaly (A), right hemisphere complex malformation with giant

heterotopia, polymicrogyria and malformed mesial temporal structures (B), and multiple cortical/ subcortical tubers in a patient with tuberous sclerosis (C). R, right.

many neuropathological subtypes of HS, but we might simplify inspection of MRI using the following criteria: reduced dimension of hippocampal formation, abnormal internal structure, loss of digitations at the level of hippocampal head, and increased T2 or FLAIR signal, usually most notable in the dentate gyrus.9,10 High field MRI, including 7T scanner images, allows refined and detailed evaluation of these structures at a subfield resolution with precise identification of hippocampal pathology and its gradient across different subfields.11 Both volume and signal abnormalities can be determined with qualitative analysis if we have in hand high quality isotropic 1 mm 3D T1-weighted image and FLAIR volumes (Figure 3.3, left panel). However, for more subtle abnormalities, postprocessing approaches might be the only means to unravel their existence and their confirmation should be interpreted in the context of the electroclinical diagnostic impression. Typical hippocampal abnormalities of MTLE usually present an anterior–posterior gradient, and their quantification in relation to normative values from healthy controls can be very useful (Figure 3.3, right panel). In particular, these maps can provide a useful 3D view to assist in tailoring the extent of resection to account for the most abnormal hippocampal section, whenever clinically warranted.

MRI FOR IDENTIFYING LESIONS IN NEOCORTICAL EPILEPSIES The major challenge for MRI in DRE is to localize lesions in the neocortex. Disruption of white matter (WM)/gray matter (GM) junction integrity, abnormalities in cortical thickness and in the gyral/sulcal folding pattern can be very difficult to identify when small lesions, such

3. Advanced MRI Imaging   | 31 

FIGURE 3.3  3T Sagittal T1-weighted and fluid-attenuated inversion recovery MRI sections illus-

trate the right hippocampal formation in a patient with right mesial temporal lobe epilepsy. Surface-based analysis maps the distribution of volume loss (blue to dark blue) and increased fluid-attenuated inversion recovery signals intensity (yellow to red) across the right subiculum and hippocampal subfields, z-scored with respect to the distribution in healthy controls.

FLAIR, fluid-attenuated inversion recovery. Source: Courtesy of Dr. Neda Bernasconi and Dr. Andrea Bernasconi, Neuroimaging of Epilepsy Laboratory, Montreal Neurological Institute.

as FCDs, still remain invisible to the expert's trained eyes. It is important to thoroughly inspect the cortical regions for FCDs, as some lesions with a clear transmantle sign might be in fact so evident that visual analysis can easily unravel them (Figures 3.1 and 3.4). Finally, attention needs to be given to the bottom of the sulci, where many FCDs can be seen more clearly.12 In the application of postprocessing and computational techniques to MRI data, it is important to correlate MRI with electroclinical information to avoid erroneous interpretation of findings that might be incidental or a consequence of a network effect of the disease, such as atrophy at a distant or connected regions to the focus.13 Comparison of individual patients’ MRI to a group of normal subjects can be done through MRI-morphometry,5 using surface-based14 and voxel-based analyses15 to quantify abnormalities in cortical folding and contrast between GM and WM. In general MRI-morphometry shows high sensitivity and localization accuracy for FCDs.15 Because in FCDs the contrast between GM and WM is blurred, surface-based metrics that follow the cortical curvature seem to be the ideal method to detect these changes.14 Among several parameters in surface-based MRI morphometry, gray and white matter junction blurring (GWB) and cortical thickness are the most informative for detection of FCDs. Typically, cortical thickness is calculated as the shortest distance between the WM surface and the GM surface at each vertex. Contrast between GM and WM can be estimated by calculating the non-normalized T1 image intensity. GWB values from individual patients can be voxel-wise compared with normal subjects using two-tailed Z-tests, generating maps of blurred GM/WM boundaries that are indicative of possible FCD topographies (Figures 3.5 and 3.6).

32   | SECTION II. NONINVASIVE EVALUATION

FIGURE 3.4  Fluid-attenuated inversion recovery (A) and coronal T1-weighted image (B) depicting

right parietal focal cortical dysplasia (FCD) with transmantle sign and blurred gray matter–white matter junction. Neuropathology of resected tissue (here in hematoxylin and eosin section) confirmed presence of balloon cells, thus indicative of FCD type IIb (C).

Source: Neuropathology Figure (C), courtesy of Dr. Marie-Christine Guiot (Department of Neuropathology, Montreal Neurological Institute).

FIGURE 3.5  Display of Zcore map of gray and white matter junction blurring (GWB) vertex-wise

comparisons between a patient with right frontal focal cortical dysplasia (lesion topography indicated by the yellow arrow) versus healthy controls. Maps were sampled to the FreeSurfer average surface space for group analysis and smoothed with a Gaussian kernel of 10 mm full width at half maximum. Contrast between gray matter (GM) and white matter (WM) was calculated as [100 × (GM − WM)]/[0.5 × (GM − WM)], where GM = 0.5 mm above and WM = 0.5 mm below the gray/ white interface. Control for multiple comparisons was performed to a vertex-wise of p 14 Hz with low initial amplitude in the right anterior insula and right inferior frontal gyrus. Ant Insula, anterior insula; Inf Fr, inferior frontal gyrus; SEEG, stereoelectroencephalography. Notes: Arrow – electrographic onset; * – clinical onset.

insula. Following the detection of the nuclear imaging abnormalities, MRI study was repeated with a 7 T magnet, which showed abnormal thickening over the right insular cortex, suggestive of focal cortical dysplasia (Figure 5.1B). The patient underwent SEEG implantation of the right temporal lobe, insula, and frontal lobe including the frontal operculum targeting areas of SPECT hyperperfusion and PET hypometabolism (Figure 5.1C and D). His SEEG recording revealed near continuous repetitive interictal discharges in the right anterior insula (RIA 1–3; RPS 1–4) and inferior frontal gyrus (RIA 7–8; Figure 5.2). Onset of his habitual seizures localized to the suspected focal cortical dysplasia (FCD) in the right anterior insula (RIA 1–3, RIA 7–9) and inferior frontal gyrus (RPS 1–3; Figure 5.3), concordant with the FDG-PET hypometabolism and the most intense SISCOM hyperperfusion abnormality.

56   | SECTION II. NONINVASIVE EVALUATION

Functional mapping was performed using bipolar stimulations at 1 Hz, 200 milliseconds pulse width, train duration 2 seconds, between 1 and 5 mA. Stimulations of the seizure onset contacts RIA 2–3 provoked an electrographic seizure with corresponding ictal semiology consisting of mirthless laughing. Stimulation of the inferior frontal gyrus (RPS 1–2) and right anterior insula (RIA 8–9) did not produce symptoms nor electrographic change. SEEG-guided radiofrequency thermocoagulation via the electrodes located in the seizure onset zone was performed (contacts RIA 1–3, RIA 7–9, and RPS 1–4). The patient remained seizure-free 3 months after the SEEG-guided radiofrequency thermocoagulation treatment.

CASE 2 A 21-year-old left-handed male had seizure onset at age 17 years. Seizures were characterized by an epigastric sensation and sense of impending doom or anxiety, with clenching of the face and bilateral upper extremities, lasting 45 to 180 seconds. He averaged two to three seizures per week, but at times seizures could occur daily. Initial investigations consisted of an MRI brain which revealed subependymal gray matter heterotopia in the anterior and posterior walls of the right occipital horn and a separate heterotopic gray matter in the occipital lobe. Scalp EEG showed focal interictal right temporal spike discharges, and right temporal interictal rhythmic delta activity. Ictal EEG of habitual seizures were of right temporal onset. STATISCOM of habitual seizure injected 14 seconds into a 55-second seizure showed right temporal hyperperfusion abnormality (Figure 5.4). Functional MRI showed left hemisphere language dominance. SEEG monitoring targeting the right limbic network and right occipital horn gray matter heterotopia (Figures 5.5 and 5.6A–D) showed abundant independent interictal activity in the right hippocampal head (RB 1–7), right amygdala (RB 1–3), heterotopia along the right

FIGURE 5.4  Case 2 right temporal statistical ictal SPECT coregistered to MRI hypermetabolism. Sta-

tistical ictal SPECT coregistered to MRI showed right temporal hyperperfusion abnormality from a patient’s typical seizure, consistent with a right temporal seizure onset. SPECT, single-photon emission computed tomography.

5. Nuclear Imaging   | 57 

FIGURE 5.5  Case 2 SEEG plan. Postimplant SEEG three-dimensional coregistration schematic show-

ing the SEEG implantation of the limbic network and right occipital horn gray matter heterotopia with tabular key. SEEG, stereoelectroencephalography.

anterior wall of the occipital horn (right anterior RAH 4–6), with occasional spread to the heterotopia along the right posterior wall of the occipital horn (right posterior RPH 4–7) and the right occipital lobe heterotopia (ROC 8–11). Habitual seizures showed ictal onset in the right hippocampus (RB 1–7; Figure 5.7). Subclinical seizures occurring in the occipital lobe heterotopic gray matter (RPH 1–2 and ROC 7–9 followed by evolution at RAH/RPH 4–7 and ROC 7–10) were recorded only after medication withdrawal. These findings were consistent with independent epileptic foci in the right anteromesial temporal lobe and the right occipital lobe heterotopic gray matter. The frequent interictal activity in the right occipital horn heterotopia was also suggestive of a potentially epileptogenic focus. As the recorded symptomatic seizures were right temporal supported by both SPECT and SEEG, the patient underwent standard right temporal lobectomy and amygdalohippocampectomy and remained without clinical seizures 11 months after surgery. This case illustrates SPECT concordance with SEEG seizure onset zone distinct from the structural abnormality of malformation of cortical development at the right occipital region.

SUMMARY Nuclear imaging allows for identification of important areas involved in seizure onset during the evaluation of drug-resistant focal epilepsy. Hyperperfusion outlined with ictal SPECT corresponds to cerebral areas of seizure involvement following isotope injection. Ictal hyperperfusion identification can be refined through SISCOM methods that subtract interictal data from ictal hyperperfusion and coregister the imaging to anatomic MRI.

58   | SECTION II. NONINVASIVE EVALUATION

FIGURE 5.6  (A–D) Case 2 SEEG coregistered to MRI. SEEG postimplantation coregistration to MRI

showing electrodes targeting the (A) right amygdala with (B) right hippocampal head, right hippocampal tail (not pictured), (C) heterotopia in the anterior wall of the right occipital horn, (D) heterotopia in the posterior wall of the right occipital horn, and right occipital heterotopia (not pictured).

RA, right amygdala; RAH, right occipital horn; RB, right hippocampal head; RPH, right occipital horn; SEEG, stereoelectroencephalography..

Additional statistical methods through STATISCOM even better define the ictal region. Surgical resection of ictal SISCOM and STATISCOM hyperperfusion may predispose to improved rates of seizure freedom in patients undergoing epilepsy surgery. In contrast to interictal testing, SPECT provides functional information based on the area of seizure activity making it a valuable addition to EEG as one of the few tools that provide ictal data. That said, it remains a resource-intensive test that requires performance during a seizure, which limits its availability. PET is an interictally-applied nuclear imaging technique that can help localize epileptogenic regions. In the most common form of PET, FDG radioisotope positron emission is measured using CT or MRI modalities. Hypometabolism interictally can be seen in areas of the brain that give rise to focal seizures. Hypometabolism is more often seen in temporal lobe epilepsy

5. Nuclear Imaging   | 59 

FIGURE 5.7  Case 2 ictal SEEG. SEEG of habitual seizures showing repetitive spike wave discharges

with progressive increasing amplitude ictal onset in the right hippocampus (RB 1–7). RA, right amygdala; RB, right hippocampal head; SEEG, stereoelectroencephalography.

than extratemporal epilepsy. PET also serves to improve prognosis counseling for temporal lobectomy. When PET data is concordant with scalp EEG and SPECT, seizure freedom rates are improved even in MRI-negative extratemporal epilepsy. PET is a common technology and is performed interictally, making it easy to use and widely available to patients. The region of PET hypometabolism is often larger than the seizure onset zone itself, and so PET is of limited value for precise seizure onset localization.

KEY REFERENCES Only

key references appear in the print edition.

The

full reference list appears in the digital product

found on http://connect.springerpub.com/content/book/978-0-8261-3693-0/part/part02/chapter/ch05

  3. O’Brien TJ, So EL, Mullan BP, et al. Subtraction SPECT co-registered to MRI improves postictal SPECT localization of seizure foci. Neurology. 1999;52:137–146. doi:10.1212/WNL.52.1.137   5. Kazemi NJ, Worrell GA, Stead SM, et al. Ictal SPECT statistical parametric mapping in temporal lobe epilepsy surgery. Neurology. 2010;74(1):70–76. doi:10.1212/WNL.0b013e3181c7da20 16. Chassoux F, Rodrigo S, Semah F, et al. FDG-PET improves surgical outcome in negative MRI Taylortype focal cortical dysplasias. Neurology. 2010;75:2168–2175. doi:10.1212/WNL.0b013e31820203a9 17. Muhlhofer W, Tan Y-L, Mueller SG, Knowlton R. MRI-negative temporal lobe epilepsy—what do we know? Epilepsia. 2017;58:727–742. doi:10.1111/epi.13699 19. Vinton AB, Carne R, Hicks RJ, et al. The extent of resection of FDG-PET hypometabolism relates to outcome of temporal lobectomy. Brain. 2007;130:548–560. doi:10.1093/brain/awl232

6 Patient Selection for Stereo EEG Jessica W. Templer and Stephan U. Schuele

KEY CONCEPTS • Invasive evaluation should be considered when: • noninvasive data is divergent, conflicting, or inconclusive • epileptogenic zone is overlapping or proximal to eloquent cortex • Secondary indications include: • to further corroborate the epileptogenic zone (i.e., epileptogenic network may extend beyond a lesion) • to provide prognostic information (e.g., to further assess likelihood of seizure freedom in a nonlesional dominant temporal lobe epilepsy case) • Stereo electroencephalography (SEEG) is typically preferred over subdural grid (SDG) implantation when there is need to record from deeper cortical structures, in nonlesional patients, for complex or multifocal lesions, bihemispheric explorations, or after a prior craniotomy. • A suspected epileptogenic zone near eloquent cortex often lends itself to a SDG evaluation. • Patient factors to consider when deciding between SEEG and SDG may include age, willingness to tolerate one versus two surgeries, ability to cooperate with a craniotomy for implantation, and patient preference. • Institutional factors to consider include neurosurgical and epileptologist experience, center experience, and whether required logistics are available for either surgical approach.

INTRODUCTION Intracranial recordings, whether obtained with SEEG, SDG, or SDG with depth electrodes, aid in identifying the epileptogenic zone and allow for delineation of appropriate surgical boundaries in patients with medically refractory epilepsy. When presurgical data is congruent with a lesion on imaging, patients can often undergo a resective surgery without obtaining further intracranial data. However, when presurgical data is divergent and/or no lesion is present on imaging, intracranial recordings are essential. In determining the best approach, either with SEEG or a SDG implantation, multiple factors should be considered, including the a priori hypothesis developed based on presurgical data, location of the suspected epileptogenic zone,

62   | SECTION II. NONINVASIVE EVALUATION

clinical case features, proximity to eloquent cortex, individual patient considerations, and institutional factors. In the majority of cases, SEEG is appropriate, allowing for identification of the epileptogenic zone with an approach bearing lower risk of morbidity and complications. Epilepsy surgery has been found to be curative in 50% to 80% of lesional temporal lobe and 40% to 60% in nonlesional temporal lobe epilepsies.1,2 Patients with extratemporal lobe epilepsy have a wider range of likelihood of seizure freedom, approximately 15% to 65%.3 Resective or thermoablative surgeries may offer a cure where medications have failed. As such, referral to an epilepsy center should be considered early if a patient continues to have seizures despite antiepileptic medications.4 Given improved outcome with early surgery, considerations for epilepsy surgery should be part of the ongoing evaluation for any patients with medically refractory focal epilepsy.5 The overarching goal of the presurgical evaluation is to identify the epileptogenic zone, the minimal area of cortex necessary to be removed to render the patient seizure-free.6 Invasive EEG monitoring guides the epilepsy surgical option(s) provided to patients. Intracranial recordings, through SEEG, SDGs, or a combination of grids and depth electrodes, aid in identifying epileptogenic tissue and appropriate surgical boundaries. By confirming an a priori hypothesis based on noninvasive data and identifying the epileptogenic zone with an invasive evaluation, a proposed surgical plan becomes a well-informed decision. The SEEG approach is designed to conceptualize the three-dimensional (3D) electroclinical seizure in a temporospatial manner.7,8 In this, the approach is fundamentally different from SDG implantations as SDGs provide a two-dimensional view of the cortex. Of patients with SDG implants, the best surgical outcomes occur in patients with cortical lesions, and in cases where SDG is utilized for functional mapping.9 Most often, SEEG will be the appropriate invasive approach in the remainder of cases (Figure 6.1). Stereo EEG implantations carry a low risk of morbidity and mortality (~1% morbidity, 0.3% mortality).10 While SDGs require a craniotomy, SEEG allows for acquisition of data via

FIGURE 6.1  Basic algorithm to determine surgical approach.

ECoG, electrocorticography; SDG, subdural grid; SEEG, stereoelectroencephalography.

6. PATIENT SELECTION FOR STEREO EEG   | 63 

a less invasive approach which is overall well-tolerated. Furthermore, SEEG provides access to recording from deep cortical structures, including depths of sulci or periventricular nodules, as well as multiple potential epileptogenic regions, both unilateral and bilateral. In lesional cases, including low grade tumors, vascular lesions, poststroke epilepsy, or posttraumatic epilepsy, a larger, complex area can be sampled to verify the hypothesis and guide resections.

DEVELOPING A HYPOTHESIS Presurgical evaluations begin with obtaining noninvasive data (i.e., scalp video EEG monitoring, MRI, neuropsychological evaluation, additional modalities as deemed appropriate) and developing a hypothesis based on this review. In lesional cases, when the neurophysiologic, clinical, and imaging data are concordant and support a single hypothesis, patients may proceed to a resection without the need for intracranial monitoring. An estimated 60% to 70% of candidates who undergo a presurgical evaluation may proceed directly to resection.11 However, invasive evaluation should be considered in the following cases:12,13 • Noninvasive data is divergent or conflicting (e.g., lesion appears to be unrelated to the electroclinical data or there are multiple potentially epileptogenic lesions) • Noninvasive data is inconclusive as is often the case in nonlesional patients • Epileptogenic zone is overlapping or proximal to eloquent cortex and functional mapping should be obtained • Secondary indications: • to further corroborate the epileptogenic zone, for example, in a patient with a cavernous malformation or glioneuronal tumor with an epileptogenic network extending beyond the lesion • to provide further prognostic information, for example, in a patient with nonlesional dominant temporal lobe epilepsy to affirm seizure onset and chances for seizure-free outcome The noninvasive data and a priori hypothesis guide the decision whether to implant with SDG or SEEG and subsequent plans for electrode placement. The importance of a well-developed hypothesis cannot be understated. The implantation scheme is developed with this hypothesis in mind, incorporating relevant data from clinical seizure semiology, scalp interictal and ictal EEG, MRI, functional MRI (fMRI), PET, and neuropsychological evaluation. In selected cases, magnetoencephalography (MEG), EEG source modeling, EEG-fMRI, and SPECT may be utilized.13 Coregistration of imaging and neurophysiologic data for magnetic source imaging, electrical source imaging or subtraction ictal SPECT coregistered to MRI is used to guide planning or to create coregistered multimodality imaging sets based on the patient’s 3D anatomy. The interpretation of SEEG results, specifically the spatial-temporal relationship of ictal events, in identifying the epileptogenic zone requires a 3D understanding of the epileptogenic cortex and surrounding areas.14 As such, incorporating additional 3D data, representing the irritative or ictal onset zone or functional anatomy and structural connectivity, often plays a more important role in planning electrode placement than in a traditional grid implantation. In many SEEG patients with a complex lesion or in nonlesional cases, these additional imaging techniques can be crucial to refine the implantation scheme.

INVASIVE MONITORING When deciding between implanting SDGs, SDGs with depth electrodes, or stereotactically implanted electrodes, multiple factors should be considered. As discussed previously, there is a fundamental difference in the approach to SDG versus SEEG, namely acquiring data in a two-dimensional versus three-dimensional space, respectively. As such, specific features and

64   | SECTION II. NONINVASIVE EVALUATION

TABLE 6.1  STEREO EEG VERSUS SUBDURAL GRID EXPLORATION CASE FEATURES

SEEG

SDG

Nonlesional cases

x

(x)

Deep cortical structures

x

Complex lesions (e.g., polymicrogyria, stroke)

x

Multiple lesions

x

Prior craniotomy/surgical failure

x

EZ near eloquent cortex

x

x

Age

>2 years old

No specific age limit

Timing of resection

Monitoring is followed by resection 6–8 weeks later

Monitoring and resection within single hospitalization

Discordant data

x

(x)

EZ, epileptogenic zone; SDG, subdural grid.

the hypothesis derived from the Phase 1 evaluation will lend itself to one approach. More specifically, discordant data, deep cortical structures, extensive structural lesions (i.e., posttraumatic epilepsy, poststroke epilepsy), multiple potential lesions, nonlesional cases, or a prior craniotomy, often favor SEEG. Whereas, when the hypothesis suggests an epileptogenic zone near eloquent cortex, SDG may be preferred (Table 6.1).

STEREO EEG VERSUS SUBDURAL GRID IMPLANTATION Tolerability

There is approximately a 1% risk of surgical complications with SEEG and less than a 0.5 % risk of mortality associated with the procedure.10 Furthermore, there is decreased perioperative pain and a shorter recovery time. A meta-analysis reviewing complications with SDGs15 reported a rate of 3.5% of patients requiring additional surgery related to complications associated with the SDG evaluation. Hemorrhagic complications associated with SEEG implantation are lower than with SDG implantation (1.0% vs. 2.4%).10,15 In addition, the rate of infectious complications is 0.8% compared to 2.3% with SDG implantation.10,15 Mortality was similar across both meta-analyses with a reported 0.3% rate of mortality.10,15

Morbidity When comparing the morbidity associated with SDG implantation to robotic SEEG, Tandon et al. reported significant differences between the two approaches.16 The patients in both groups were similar in terms of gender, age, and duration of epilepsy. The overall number of recording days were similar in both groups. The mean number of electrodes were significantly greater in SEEG cases (186.9 vs. 114.3 in SDG). However, time in the operating room was longer and in addition, major complications, transfusion requirements, and higher postoperative narcotic use were more common in the SDG group.16

Outcome Patient epilepsy characteristics are often fundamentally different between patients undergoing SEEG compared to SDG; therefore it is difficult to compare outcome between the two patient

6. PATIENT SELECTION FOR STEREO EEG   | 65 

populations. In fact, when SEEG and SDG are readily available at an epilepsy center, patients undergoing SDG are more often deemed likely resection candidates prior to the implant. The majority of epileptologists would not pursue a craniotomy for invasive monitoring if resection was not a likely outcome. This was apparent in Tandon et al.’s retrospective analysis; 91.4% of patients evaluated with SDG and 74.4% with SEEG went on to undergo resective or ablative surgery (p 250 Hz) depending on the spectral frequency have been described in ICEEG recordings during wakefulness and sleep, both in cortical and mesial temporal structures. HFO can be visually detected by filtering the EEG signal (ICEEG or scalp EEG) with a high-pass filter set at 80 Hz, by reducing the time window to around 1 second, and by increasing the gain. They have generated much interest over the past years given their physiological role in cognitive functions, including memory consolidation during sleep, but also given their recent description as a biomarker for epileptogenicity.80–83 Several characteristics have been proposed to differentiate physiological from pathological HFO, with a large overlap however between these two entities. Physiological ripples are predominant (rate >0.2/minute) in the occipital cortex, temporal lobe (medial and basal region, transverse temporal gyrus), and pre- and postcentral gyri.3 They predominate during NREM sleep and show a temporal coupling with slow wave oscillations, with an occurrence after the peak of the down state.3,84 For further information on HFO, the reader is referred to Chapter 26 in this book.

24. PHYSIOLOGICAL ACTIVITY RECORDED WITH INTRACRANIAL EEG   | 313 

PHYSIOLOGICAL ACTIVITY DURING TRANSITIONAL STATES A major contribution of ICEEG studies in sleep investigation has concerned transitional states (from wakefulness to sleep and from sleep to wakefulness), as well as transient, and mostly focal intrusions of a vigilance state into another one, such as arousals. Indeed, ICEEG recordings in humans have provided evidence that electrophysiological features of sleep and wakefulness could coexist at the same time in different brain areas.85

Sleep Onset The demonstration that the process of falling asleep did not involve the whole brain at the same time was made by Magnin et al,86 who quantified findings of Caderas et al.55 In this work, the authors used the dimension of activation (DA) which is global measure of the signal complexity allowing signal quantification and thus discrimination between vigilance stages. They reported that thalamic deactivation (decrease in DA) at sleep onset most often preceded that of the cortex by several minutes and that an asynchrony was also observed within the cortex.86 Such delay was also found in the hippocampus, where spindles not only precede neocortical spindles, but may occur several minutes before scalp-detected sleep onset.55,87

Awakening Imaging and EEG studies have suggested that awakening was also a progressive process associated with the sequential reactivation of brain structures underlying sleep inertia phenomena.88 In rodents, awakening from sleep is characterized by the presence, during several minutes, of remaining local slow waves associated with “OFF-periods.”89 Few SEEG studies have specifically studied the sleep-to-wake transition except in pathological situations such as parasomnia.90 Regarding physiological awakening from NREM sleep, Magnin et al. did not report thalamo-cortical nor intra-cortical asynchrony; however, except for the thalamus (mostly the medial pulvinar), no other subcortical region was investigated and it is possible that DA may not explore short-lasting events such as “OFF periods”.86

Micro-Arousals Arousal are defined as brief (3–15 seconds) activations during sleep, associated with an abrupt shift of EEG spectrum toward higher frequencies and, for REM sleep arousals, with an increase in muscle tone.91 Such activations can be observed locally during NREM sleep in the motor cortex (blockage of slow waves and increase in alpha–beta frequencies) whereas other areas, especially the dorsolateral prefrontal cortex, exhibit an increase in slow wave activity.92 They occur also during REM sleep, allowing initiation of a motor response without global awakening.93 Stereotypical arousals at the thalamic level are associated with heterogeneous patterns of cortical arousals mainly depending on the cortical area considered and the ongoing sleep stage.94 Interestingly, the SEEG signal during arousals differs from that recorded during awakenings, suggesting that the shift between sleep and wake may actually be progressive.94

Micro-Sleep The presence of local cortical slow waves and “OFF periods” after extended wakefulness in rats, correlated with impaired performance, suggests that intrusion of sleep features during wakefulness may represent a sleep homeostatic phenomenon.95 Such “micro-sleep” could constitute the electrophysiological correlate of sleepiness. Accordingly, HD-EEG recordings have provided evidence of local aspects of sleep during wakefulness in children.96 Only one study with ECoG

314   | SECTION VII. INTERPRETATION

has reported that the detection of such local inactive micro-states, associated with decrease in theta–alpha and increase in delta frequencies, increased with the time spent awake.97

SUMMARY ICEEG recordings have allowed not only confirmation of the presence of most previously scalp EEG-defined features of sleep, but also advancement of their description and presentation of hypotheses regarding their functional significance. The most important contributions of ICEEG recordings concern detailed analyses of topography and the investigation of deep structures. Thus, our understanding of sleep has evolved over the last 10 years with the demonstration of spatial and temporal inhomogeneities suggesting that vigilance states evolve along a continuum and are regulated locally. Moreover, intracranial recordings including investigation of mesio-temporal structures during sleep have provided evidence that within and across area coupling of sleep oscillations support cognitive functions of sleep. However, several fields remain incompletely explored, and could benefit from larger database analysis, which could allow overcoming the spatial sampling bias of SEEG, and translational approaches, especially regarding subcortical structures recordings in movement disorders, simultaneous multimodal explorations including functional imaging, and age-specific investigation of maturational aspects of sleep.

KEY REFERENCES Only

key references appear in the print edition.

The

full reference list appears in the digital product

found on http://connect.springerpub.com/content/book/978-0-8261-3693-0/part/part07/chapter/ch24

  1. Frauscher B, von Ellenrieder N, Zelmann R, et al. Atlas of the normal intracranial electroencephalogram: neurophysiological awake activity in different cortical areas. Brain. 2018;141(4):1130–1144. doi:10.1093/brain/awy035   2. von Ellenrieder N, Gotman J, Zelmann R, et al. How the human brain sleeps: direct cortical recordings of normal brain activity. Ann Neurol. 2020;87(2):289–301. doi:10.1002/ana.25651 18. Halgren M, Ulbert I, Bastuji H, et al. The generation and propagation of the human alpha rhythm. Proc Natl Acad Sci U S A. 2019;116(47):23772–23782. doi:10.1073/pnas.1913092116 33. Staresina BP, Bergmann TO, Bonnefond M, et al. Hierarchical nesting of slow oscillations, spindles and ripples in the human hippocampus during sleep. Nat Neurosci. 2015;18(11):1679–1686. doi:10.1038/nn.4119 36. Csercsa R, Dombovári B, Fabó D, et al. Laminar analysis of slow wave activity in humans. Brain. 2010;133(9):2814–2829. doi:10.1093/brain/awq169 56. Andrillon T, Nir Y, Staba RJ, et al. Sleep spindles in humans: insights from intracranial EEG and unit recordings. J Neurosci. 2011;31(49):17821–17834. doi:10.1523/JNEUROSCI.2604-11.2011 63. De Carli F, Proserpio P, Morrone E, et al. Activation of the motor cortex during phasic rapid eye movement sleep. Ann Neurol. 2016;79(2):326–330. doi:10.1002/ana.24556 86. Magnin M, Rey M, Bastuji H, et al. Thalamic deactivation at sleep onset precedes that of the cerebral cortex in humans. Proc Natl Acad Sci U S A. 2010;107(8):3829–3833. doi:10.1073/ pnas.0909710107 94. Peter-Derex L, Magnin M, Bastuji H. Heterogeneity of arousals in human sleep: a stereo-electroencephalographic study. Neuroimage. 2015;123:229–244. doi:10.1016/j.neuroimage.2015.07.057

25 Interpretation of Interictal Epileptiform Discharges Shasha Wu, Naoum P. Issa, Sandra Rose, and James X. Tao

KEY CONCEPTS • Interictal epileptiform discharges (IEDs) can help in localizing the seizure onset zone (SOZ) and may indicate the extent of the epileptogenic zone (EZ). • The spatial distribution of the irritative zone (IZ) is highly variable among patients, and within an individual the boundaries of the IZ may change over the course of the disease. • The IEDs with the highest amplitude, the earliest peak (leading spike), and the highest frequency are more likely to be found within the SOZ. • Combining the pattern of IEDs with high frequency oscillations (HFOs) and functional MRI (fMRI) may increase the value of IEDs in seizure localization. • Specific patterns of intracerebral IEDs have been reported to correlate with specific pathologies.

INTRODUCTION The goal of intracerebral depth electrode EEG recording is to record seizures and the propagation pattern of the seizures in order to delineate the extent of the EZ. EZ is defined as the “area of cortex that is necessary and sufficient for initiating seizures and whose removal (or disconnection) is necessary for the complete abolition of seizures.”1 Historically, the concept of the EZ was proposed based on the recording of IEDs and the importance of IEDs was emphasized by Penfield more than a half-century ago: “we feel that sporadic spikes are the most reliable index of a primary epileptic discharge of the cortex.”2 The IEDs were thought to represent pathological alterations in normal cellular excitability and synchronization.3 In modern epilepsy presurgical evaluation, the IEDs define the IZ.4 Although the relationship between IZ and the SOZ is not completely understood, IEDs are undoubtedly important components to be analyzed in intracranial recording to confirm the SOZ and to guide the extent of the surgery.5 This chapter focuses on the interpretation of IEDs recorded by chronically implanted depth electrodes and discusses its role in the epilepsy presurgical evaluation.

316   | SECTION VII. INTERPRETATION

IDENTIFICATION OF INTERICTAL EPILEPTIFORM DISCHARGES ON STEREO EEG Compared to IEDs recorded on the scalp, intracerebrally recorded IEDs are higher in amplitude, more frequent, have more complex and sharper configuration, and are often multifocal.

MORPHOLOGY IEDs recorded in stereoelectroencephalography (SEEG) can take on a variety of morphologies and durations. However, like IEDs on scalp EEG, they usually have a “sharp” component and should be clearly distinct from background activity (Figure 25.1). Sharp waves have a pointed peak and a duration of 70 to 200 milliseconds, and the ascending phase of the sharp wave is usually steeper than the descending phase. Spikes are similar to sharp waves but have a duration between 20 and 70 milliseconds. Most spikes and sharp waves are followed by a slow wave, but the terms “sharp-and-slow-wave complex” and “spike-and-slow-wave complex” classically refer to discharges in which the slow wave is of higher amplitude than the sharp wave or spike. Multiple spike-and-slow-wave complexes are similar to spike-and-slow-wave complexes but have two or more spikes associated with one or more slow waves.6 Other IED patterns, including bursts of fast spikes, repetitive spikes, and so on, have also been described

FIGURE 25.1  Different patterns of interictal epileptiform discharges recorded in human partial sei-

zures with stereotactic EEG. (A) Periodic sharp waves. (B) Isolated spike. (C) Polyspikes. (D) Spike with superimposed high frequency oscillations. All the SEEG are displayed with band-pass filter settings of 1 to 100 Hz. SEEG, stereoelectroencephalography.

25. INTERPRETATION OF INTERICTAL EPILEPTIFORM DISCHARGES   | 317 

in the SEEG recording.7,8 Due to the large variety of IED patterns even in a single patient, it has been assumed that different types of IEDs are generated by distinct neurobiological mechanisms and pathologies and have differing roles in ictogenesis.9,10

AMPLITUDE The amplitude of IEDs on scalp EEG varies but is typically 50 to 200 µV; IEDs recorded intracranially are usually 8 to 20 times larger than those on scalp EEG11 and in the range of several hundred microvolts to millivolts. The larger amplitude of intracranial IEDs is explained by the physical characteristics of volume conduction of electrical signals. The amplitude of a voltage dipole falls with the square of distance;12 because intracerebral electrodes lie directly on or within the cortex, they are more likely closer to the spike generator than are scalp electrodes and can detect the discharge before its amplitude has decayed with distance. In addition, dura, skull and scalp soft tissue attenuate the intracranial potentials, leading to much smaller amplitude IEDs on scalp EEG.13 While not affecting the amplitude of IEDs, intracranially recorded potentials are also less likely to be affected by muscle artifact, resulting in a larger signal-to-noise ratio compared to scalp EEG. Numerous studies have reported that the IEDs with the highest amplitude, the earliest peak (leading spike), and the highest frequency are often closer to or colocalize with the SOZ.14–16

POLARITY EEG electrodes predominantly record the voltage changes produced by pyramidal neurons, which are oriented perpendicularly to the pial surface.12 Because of the elongated dendritic tree of pyramidal neurons, the voltage changes associated with IEDs are usually polarized, (with a negative maximum near the pial surface and a positive maximum near the grey–white boundary). As a result, IEDs recorded from the pial surface of the cortex, as would be the case for subdural grid electrodes or scalp EEG electrodes, have a main negative sharp component, followed by a small positive deflection and often a large negative slow-wave. The initial negative deflection correlates with an excitatory postsynaptic depolarization from superficial laminae of the cortex, and the slow-wave is related to a long-lasting hyperpolarization (inhibition).17,18 A low-voltage positive wave preceding the main negative sharp wave is commonly seen and is thought to correspond to the excitatory postsynaptic potential synchronization over deep cortical lamina.19 It is not uncommon that a distinct low-voltage negative wave occurs before the low-voltage positive wave, which has been named the “n wave” by Serafini (Figure 25.2).20 The n-wave is hypothesized to be generated by sequential recruitment of multiple distinct epileptic micro-foci over superficial cortical laminae and may be associated with chronic epilepsy.20 Since SEEG electrodes can be placed in either superficial or deeper cortical layers, the IEDs recorded with SEEG can have either a positive or a negative maximum. If an SEEG electrode array spans the generator of the IED, a reversal of polarity might be visible across the array. A shifting of polarity can also be seen in the same recording contact.5 Figure 25.3 shows an example of polarity change within one contact and across several contacts in the depth electrodes recorded from a patient with right mesial temporal epilepsy.

EVOLUTION IEDs can propagate through cerebral circuitry, producing a temporally and spatially evolving pattern. One common way of following IED propagation is by comparing the latency of peaks across different channels. However, the midpoint of the ascending negative phase (rather than the peak) is recommended to be used for source localization.21,22 It has been suggested that

318   | SECTION VII. INTERPRETATION

FIGURE 25.2  The illustration of an n-wave. The very early onset of an interictal epileptiform dis-

charge consists of a low-voltage negative polarity waveform (1) that precedes a small positive prepotential (2) followed by the common high-amplitude negative sharp wave (3), then a positive deflection (4) and a negative slowing wave (5). Source: Figure and legend adapted from Serafini R. Similarities and differences between the interictal epileptiform discharges of green-spikes and red-spikes zones of human neocortex. Clin Neurophysiol. 2019;130(3):396– 405. doi:10.1016/j.clinph.2018.12.011

the area generating the earliest IED peak, or the “leading region,” has a high correlation to the location of epileptogenic brain tissue.23 Hufnagel et al. retrospectively studied intracranial EEG recording from 32 patients (a combination of subdural electrodes and depth electrodes) and found a good correlation between the SOZ and the location with the earliest IED peak.15 The authors concluded that the relative latency of the spike could be a more reliable marker for the SOZ than the amplitude or frequency of spiking.

INTERICTAL EPILEPTIFORM ACTIVITY, SEIZURE ONSET ZONE, AND THE EPILEPTOGENIC ZONE The EZ is composed of a network of interconnected structures able to initiate seizures and generate epileptiform discharges during the interictal period.24 The current practice for identification of the EZ is by detecting the SOZ on scalp or intracranial EEG. In real-world situations, IEDs alone, without captured seizures, are usually insufficient to decide what piece of tissue should be resected, but it is commonly accepted that IEDs reflect transient activation of epileptogenic networks.25 Complete resection of the SOZ does not always lead to seizure freedom, meaning that the EZ is often more extensive that the SOZ. Some have suggested that IEDs could help define the extent of the epileptogenic network.26 The distinction between spikes associated with the SOZ and those outside the SOZ was formalized in the nomenclature of “red spikes” and “green spikes.” Red spikes were proposed to describe the IEDs originating from brain area generating both IEDs and seizures. Green spikes were proposed to describe IEDs recorded only from nonseizure-generating tissue in which the epileptogenic threshold was lowered due to synaptic connections with the primarily pathological cortex, or as a result of neuronal propagation from the primary epileptic area.27–29 Together the regions that generate red and green spikes constitute the IZ.30 Studies have tried to define the distinguishing characteristics of red spikes and green spikes. Serafini and Loeb demonstrated that the IEDs at the center of the epileptogenic focus exhibited a prominent sharp wave while IEDs at the periphery of the epileptogenic focus exhibited a more pronounced slow wave.31 Similarly, the IEDs in the green spike–generating cortex (green spike zone) exhibited more pronounced slow waves than the red spike–generating cortex (red spike zone) and supported more neurophysiological peripheral inhibition in the nonepileptogenic tissue.32

25. INTERPRETATION OF INTERICTAL EPILEPTIFORM DISCHARGES   | 319 

FIGURE 25.3  The interictal sharp waves with polarity reversal are recorded from a hippocampal

depth electrode in a patient with right mesial temporal lobe epilepsy. The polarity of the interictal epileptiform discharges reverses from negative (*) to positive in the same electrode right hippocampal depth (RHD) 2 (**) and across electrodes RHD 2 to 10 (#). Total 12 contacts, contact 1 is the deepest. The three-dimensional reconstruction of the intracranial electrode locations with the patient’s MRI is shown inset. Red structure is the amygdala, yellow structure is the hippocampus. RHD 1, 2, and 3 are in the amygdala. RHD, right hippocampal depth.

IEDs in SEEG recording can occur in isolation or in short bursts. Several groups have reported that the spikes with the highest frequency were more likely to colocalize with the SOZ than spikes with lower frequency.33,34 For example, a good correlation was found between the area with frequent spikes and the EZ in patients with focal cortical dysplasia (FCD).26 However, identifying a region with high-frequency spiking did not always imply a good outcome: 1 year after surgery for temporal lobe epilepsy, only four of 14 (28.6%) patients with frequent spikes (>60 spikes/hour) were seizure-free compared to 33 of 41 (80.5%) patients with infrequent

320   | SECTION VII. INTERPRETATION

(