Epileptic Seizures and the Eeg: Measurement, Models, Detection and Prediction

Analysis of medical data using engineering tools is a rapidly growing area, both in research and in industry, yet few texts exist that address the problem from an interdisciplinary perspective. Epileptic Seizures and the EEG: Measurement, Models, Detection and Prediction brings together biology and engineering practices and identifies the aspects of the field that are most important to the analysis of epilepsy.
Analysis of EEG records
The book begins by summarizing the physiology and the fundamental ideas behind the measurement, analysis and modeling of the epileptic brain. It introduces the EEG as a measured signal and explains its use in the study of epilepsy. Next, it provides an explanation of the type of brain activity likely to register in EEG measurements, offering quantitative analysis of the populations of neurons that contribute to both scalp and cortical EEG and discussing the limitations and effects that choices made in the recording process have on the data. The book provides an overview of how these EEG records are and have been analyzed in the past, concentrating on the mathematics relevant to the problem of classification of EEG. The authors use these extracted features to differentiate between or classify inter-seizure, pre-seizure and seizure EEG.
The challenge of seizure prediction

The book focuses on the problem of seizure detection and surveys the physiologically based dynamic models of brain activity. Finally, the book addresses the fundamental question: can seizures be predicted? Through analysis of epileptic activity spanning from 3 hours to 25 years, it is proposed that seizures may be predictable, but the amount of data required is greater than previously thought. Based on the authors’ extensive research, the book concludes by exploring a range of future possibilities in seizure prediction.

Andrea Varsavsky and Iven Mareels are with The University of Melbourne in Victoria, Australia. Mark Cook is with St. Vincent’s Hospital and the University of Melbourne in Victoria, Australia.

Introduction The Brain and Epilepsy Micro-scopic Dynamics: Single Neurons Meso/Macro-scopic Dynamics: Neural Networks Neurotransmitters and Neuromodulators Epilepsy - A Malfunctioning Brain Diagnosis and Treatment of Epilepsy The EEG - A Recording of the Brain The Normal EEG The Epileptic EEG Detecting Changes in the EEG Dynamics of the Brain Micro- and Macro-scopic models Dynamic Models of Epilepsy Stochasticity in Neural Systems EEG Generation and Measurement Principles of Bioelectric Phenomena A Foreword On Notation From Single Charges to Equivalent Dipoles Equivalent Current Dipoles Macro-scopic Mean Fields - Homogeneous Media Macro-scopic Mean Fields - Inhomogeneous Media Current Sources in Biological Tissue Synaptic Structure and Current Dipoles Spatial Integration Temporal Integration Volume Conducting Properties of the Head Head Geometry Capacitive Effects of Tissue Estimating Conductivities The EEG: A Macro-scopic View of the Brain EEG Measurement EEG Dynamics Epilepsy and the EEG Appendix A: Units of Electric Quantities Appendix B: Volume Conductor Boundary Conditions Appendix:C: Capacitance in RC Circuits Signal Processing in EEG Analysis Mathematical Representation of the EEG Preprocessing Feature Extraction Time Domain Analysis Frequency Domain Analysis Time-Frequency Analysis Nonlinear Analysis Detection and Prediction of Seizures in Literature Classifying the EEG Types of Classifiers Association Rules Artiificial Neural Networks Support Vector Machines Expert System Processing Decisions Spatio-Temporal Context Patient Specificity Seizure Detection The Problem of Seizure Detection The EEG Database Performance Evaluation Metrics Evaluation of Classification Methods Feature Extraction ANN Training and Testing SVM Training and Testing Results and Comparisons Evaluation of Patient Un-specific Seizure Detectors Algorithm 1: Monitor Algorithm 2: CNet Algorithm 3: Reveal Algorithm 4: Saab Comparisons and Conclusions Evaluation of Onset Seizure Detectors Feature Extraction Results and Comparisons Modeling for Epilepsy Physiological Parameters of Neural Models Parameters in Single Neurons Parameters in Networks of Neurons Micro-scopic (Statistical) Models Model Summary Validation and Limitations Meso-scopic (Phenomenological) Models Model Summary Analysis: Linearization, Stability and Instability Validation and Limitations: Rhythms in the EEG Relationship to Micro-scopic Models Macro-scopic Models (Future Outlook) Practical Use of Models Epileptic Seizure Generation Limitations of the EEG Appendix A: Physiological Parameters and Notation Appendix B: Summary of IF Model Appendix C: Summary of Phenomenological Model On the Predictability of Seizures Predictability - Terminology Made Clear How to Estimate LRD Example Distributions Computing α Simulations Results Seizure Frequency Dataset Analysis - Estimation of α Memory and Predictability of Seizures Concluding Remarks