Eeg Signal Processing and Machine Learning

ISBN-13: 9781119386940 / Angielski / Twarda / 2021 / 752 str.

Saeid Sanei

Opis

The book aims at describing new techniques and outcomes in electroencephalogram (EEG) research mainly in analysis, processing and decision making about various brain states, abnormalities, and disorders by means of advance signal processing and machine learning techniques respectively.

Spis treści

Preface to the Second Edition xviiPreface to the First Edition xxiList of Abbreviations xxiii1 Introduction to Electroencephalography 11.1 Introduction 11.2 History 21.3 Neural Activities 51.4 Action Potentials 61.5 EEG Generation 81.6 The Brain as a Network 121.7 Summary 12References 132 EEG Waveforms 152.1 Brain Rhythms 152.2 EEG Recording and Measurement 182.2.1 Conventional Electrode Positioning 212.2.2 Unconventional and Special Purpose EEG Recording Systems 242.2.3 Invasive Recording of Brain Potentials 262.2.4 Conditioning the Signals 272.3 Sleep 282.4 Mental Fatigue 302.5 Emotions 302.6 Neurodevelopmental Disorders 312.7 Abnormal EEG Patterns 322.8 Ageing 332.9 Mental Disorders 342.9.1 Dementia 342.9.2 Epileptic Seizure and Nonepileptic Attacks 352.9.3 Psychiatric Disorders 392.9.4 External Effects 402.10 Summary 41References 423 EEG Signal Modelling 473.1 Introduction 473.2 Physiological Modelling of EEG Generation 473.2.1 Integrate-and-Fire Models 493.2.2 Phase-Coupled Models 493.2.3 Hodgkin-Huxley Model 513.2.4 Morris-Lecar Model 543.3 Generating EEG Signals Based on Modelling the Neuronal Activities 573.4 Mathematical Models Derived Directly from the EEG Signals 613.4.1 Linear Models 613.4.1.1 Prediction Method 613.4.1.2 Prony's Method 623.4.2 Nonlinear Modelling 643.4.3 Gaussian Mixture Model 663.5 Electronic Models 673.5.1 Models Describing the Function of the Membrane 673.5.1.1 Lewis Membrane Model 683.5.1.2 Roy Membrane Model 683.5.2 Models Describing the Function of a Neuron 683.5.2.1 Lewis Neuron Model 683.5.2.2 The Harmon Neuron Model 713.5.3 A Model Describing the Propagation of the Action Pulse in an Axon 723.5.4 Integrated Circuit Realizations 723.6 Dynamic Modelling of Neuron Action Potential Threshold 723.7 Summary 73References 734 Fundamentals of EEG Signal Processing 774.1 Introduction 774.2 Nonlinearity of the Medium 784.3 Nonstationarity 794.4 Signal Segmentation 804.5 Signal Transforms and Joint Time-Frequency Analysis 834.5.1 Wavelet Transform 874.5.1.1 Continuous Wavelet Transform 874.5.1.2 Examples of Continuous Wavelets 894.5.1.3 Discrete-Time Wavelet Transform 894.5.1.4 Multiresolution Analysis 904.5.1.5 Wavelet Transform Using Fourier Transform 934.5.1.6 Reconstruction 944.5.2 Synchro-Squeezed Wavelet Transform 954.5.3 Ambiguity Function and the Wigner-Ville Distribution 964.6 Empirical Mode Decomposition 1004.7 Coherency, Multivariate Autoregressive Modelling, and Directed Transfer Function 1014.8 Filtering and Denoising 1044.9 Principal Component Analysis 1074.9.1 Singular Value Decomposition 1084.10 Summary 110References 1105 EEG Signal Decomposition 1155.1 Introduction 1155.2 Singular Spectrum Analysis 1155.2.1 Decomposition 1165.2.2 Reconstruction 1175.3 Multichannel EEG Decomposition 1185.3.1 Independent Component Analysis 1185.3.2 Instantaneous BSS 1225.3.3 Convolutive BSS 1265.3.3.1 General Applications 1275.3.3.2 Application of Convolutive BSS to EEG 1285.4 Sparse Component Analysis 1295.4.1 Standard Algorithms for Sparse Source Recovery 1305.4.1.1 Greedy-Based Solution 1305.4.1.2 Relaxation-Based Solution 1315.4.2 k-Sparse Mixtures 1315.5 Nonlinear BSS 1335.6 Constrained BSS 1345.7 Application of Constrained BSS; Example 1355.8 Multiway EEG Decompositions 1365.8.1 Tensor Factorization for BSS 1395.8.2 Solving BSS of Nonstationary Sources Using Tensor Factorization 1435.9 Tensor Factorization for Underdetermined Source Separation 1495.10 Tensor Factorization for Separation of Convolutive Mixtures in the Time Domain 1535.11 Separation of Correlated Sources via Tensor Factorization 1535.12 Common Component Analysis 1545.13 Canonical Correlation Analysis 1545.14 Summary 155References 1556 Chaos and Dynamical Analysis 1656.1 Introduction to Chaos and Dynamical Systems 1656.2 Entropy 1666.3 Kolmogorov Entropy 1666.4 Multiscale Fluctuation-Based Dispersion Entropy 1676.5 Lyapunov Exponents 1676.6 Plotting the Attractor Dimensions from Time Series 1696.7 Estimation of Lyapunov Exponents from Time Series 1696.7.1 Optimum Time Delay 1726.7.2 Optimum Embedding Dimension 1726.8 Approximate Entropy 1736.9 Using Prediction Order 1746.10 Summary 175References 1757 Machine Learning for EEG Analysis 1777.1 Introduction 1777.2 Clustering Approaches 1817.2.1 k-Means Clustering Algorithm 1817.2.2 Iterative Self-Organizing Data Analysis Technique 1837.2.3 Gap Statistics 1837.2.4 Density-Based Clustering 1847.2.5 Affinity-Based Clustering 1847.2.6 Deep Clustering 1847.2.7 Semi-Supervised Clustering 1857.2.7.1 Basic Semi-Supervised Techniques 1857.2.7.2 Deep Semi-Supervised Techniques 1867.2.8 Fuzzy Clustering 1867.3 Classification Algorithms 1877.3.1 Decision Trees 1887.3.2 Random Forest 1897.3.3 Linear Discriminant Analysis 1907.3.4 Support Vector Machines 1917.3.5 k-Nearest Neighbour 1997.3.6 Gaussian Mixture Model 2007.3.7 Logistic Regression 2007.3.8 Reinforcement Learning 2017.3.9 Artificial Neural Networks 2017.3.9.1 Deep Neural Networks 2037.3.9.2 Convolutional Neural Networks 2057.3.9.3 Autoencoders 2077.3.9.4 Variational Autoencoder 2087.3.9.5 Recent DNN Approaches 2097.3.9.6 Spike Neural Networks 2107.3.9.7 Applications of DNNs to EEG 2127.3.10 Gaussian Processes 2127.3.11 Neural Processes 2137.3.12 Graph Convolutional Networks 2137.3.13 Naïve Bayes Classifier 2137.3.14 Hidden Markov Model 2147.3.14.1 Forward Algorithm 2167.3.14.2 Backward Algorithm 2167.3.14.3 HMM Design 2167.4 Common Spatial Patterns 2187.5 Summary 222References 2238 Brain Connectivity and Its Applications 2358.1 Introduction 2358.2 Connectivity through Coherency 2388.3 Phase-Slope Index 2408.4 Multivariate Directionality Estimation 2408.4.1 Directed Transfer Function 2418.4.2 Direct DTF 2428.4.3 Partial Directed Coherence 2438.5 Modelling the Connectivity by Structural Equation Modelling 2438.6 Stockwell Time-Frequency Transform for Connectivity Estimation 2468.7 Inter-Subject EEG Connectivity 2478.7.1 Objectives 2478.7.2 Technological Relevance 2478.8 State-Space Model for Estimation of Cortical Interactions 2498.9 Application of Cooperative Adaptive Filters 2518.9.1 Use of Cooperative Kalman Filter 2538.9.2 Task-Related Adaptive Connectivity 2548.9.3 Diffusion Adaptation 2558.9.4 Brain Connectivity for Cooperative Adaptation 2568.9.5 Other Applications of Cooperative Learning and Brain Connectivity Estimation 2578.10 Graph Representation of Brain Connectivity 2588.11 Tensor Factorization Approach 2598.12 Summary 262References 2639 Event-Related Brain Responses 2699.1 Introduction 2699.2 ERP Generation and Types 2699.2.1 P300 and its Subcomponents 2739.3 Detection, Separation, and Classification of P300 Signals 2749.3.1 Using ICA 2759.3.2 Estimation of Single-Trial Brain Responses by Modelling the ERP Waveforms 2779.3.3 ERP Source Tracking in Time 2789.3.4 Time-Frequency Domain Analysis 2809.3.5 Application of Kalman Filter 2849.3.6 Particle Filtering and its Application to ERP Tracking 2869.3.7 Variational Bayes Method 2919.3.8 Prony's Approach for Detection of P300 Signals 2939.3.9 Adaptive Time-Frequency Methods 2979.4 Brain Activity Assessment Using ERP 2989.5 Application of P300 to BCI 2999.6 Summary 300References 30110 Localization of Brain Sources 30710.1 Introduction 30710.2 General Approaches to Source Localization 30810.2.1 Dipole Assumption 30910.3 Head Model 31110.4 Most Popular Brain Source Localization Approaches 31310.4.1 EEG Source Localization Using Independent Component Analysis 31310.4.2 MUSIC Algorithm 31310.4.3 LORETA Algorithm 31710.4.4 FOCUSS Algorithm 31810.4.5 Standardized LORETA 31910.4.6 Other Weighted Minimum Norm Solutions 32010.4.7 Evaluation Indices 32310.4.8 Joint ICA-LORETA Approach 32310.5 Forward Solutions to the Localization Problem 32510.5.1 Partially Constrained BSS Method 32510.5.2 Constrained Least-Squares Method for Localization of P3a and P3b 32610.5.3 Spatial Notch Filtering Approach 32810.6 The Methods Based on Source Tracking 33310.6.1 Deflation Beamforming Approach for EEG/MEG Multiple Source Localization 33310.6.2 Hybrid Beamforming - Particle Filtering 33610.7 Determination of the Number of Sources from the EEG/MEG Signals 33710.8 Other Hybrid Methods 34010.9 Application of Machine Learning for EEG/MEG Source Localization 34010.10 Summary 342References 34311 Epileptic Seizure Prediction, Detection, and Localization 35111.1 Introduction 35111.2 Seizure Detection 35711.2.1 Adult Seizure Detection from EEGs 35711.2.2 Detection of Neonatal Seizure 36311.3 Chaotic Behaviour of Seizure EEG 36611.4 Seizure Detection from Brain Connectivity 36911.5 Prediction of Seizure Onset from EEG 36911.6 Intracranial and Joint Scalp-Intracranial Recordings for IED Detection 38411.6.1 Introduction to IED 38411.6.2 iEED-Times IED Detection from Scalp EEG 38611.6.3 A Multiview Approach to IED Detection 39111.6.4 Coupled Dictionary Learning for IED Detection 39111.6.5 A Deep Learning Approach to IED Detection 39211.7 Fusion of EEG-fMRI Data for Seizure Prediction 39611.8 Summary 398References 39912 Sleep Recognition, Scoring, and Abnormalities 40712.1 Introduction 40712.1.1 Definition of Sleep 40712.1.2 Sleep Disorder 40812.2 Stages of Sleep 40912.2.1 NREM Sleep 40912.2.2 REM Sleep 41112.3 The Influence of Circadian Rhythms 41412.4 Sleep Deprivation 41512.5 Psychological Effects 41612.6 EEG Sleep Analysis and Scoring 41612.6.1 Detection of the Rhythmic Waveforms and Spindles Employing Blind Source Separation 41612.6.2 Time-Frequency Analysis of Sleep EEG Using Matching Pursuit 41712.6.3 Detection of Normal Rhythms and Spindles Using Higher-Order Statistics 42112.6.4 Sleep Scoring Using Tensor Factorization 42312.6.5 Application of Neural Networks 42512.6.6 Model-Based Analysis 42612.7 Detection and Monitoring of Brain Abnormalities during Sleep by EEG and Multimodal PSG Analysis 42812.7.1 Analysis of Sleep Apnoea 42812.7.2 EEG and Fibromyalgia Syndrome 43112.7.3 Sleep Disorders of Neonates 43112.8 Dreams and Nightmares 43212.9 EEG and Consciousness 43312.10 Functional Brain Connectivity for Sleep Analysis 43312.11 Summary 434References 43513 EEG-Based Mental Fatigue Monitoring 44113.1 Introduction 44113.2 Feature-Based Machine Learning Approaches 44313.2.1 Hidden Markov Model Application 44313.2.2 Kernel Principal Component Analysis and Hidden Markov Model 44413.2.3 Regression-Based Fatigue Estimation 44413.2.4 Regularized Regression 44513.2.5 Other Feature-Based Approaches 44513.3 Measurement of Brain Synchronization and Coherency 44613.3.1 Linear Measure of Synchronization 44613.3.2 Nonlinear Measure of Synchronization 44813.4 Evaluation of ERP for Mental Fatigue 45113.5 Separation of P3a and P3b 45713.6 A Hybrid EEG-ERP-Based Method for Fatigue Analysis Using an Auditory Paradigm 46313.7 Assessing Mental Fatigue by Measuring Functional Connectivity 46513.8 Deep Learning Approaches for Fatigue Evaluation 47213.9 Summary 474References 47414 EEG-Based Emotion Recognition and Classification 47914.1 Introduction 47914.1.1 Theories and Emotion Classification 48014.1.2 The Physiological Effects of Emotions 48214.1.3 Psychology and Psychophysiology of Emotion 48514.1.4 Emotion Regulation 48714.1.4.1 Agency and Intentionality 49014.1.4.2 Norm Violation 49014.1.4.3 Guilt 49114.1.4.4 Shame 49114.1.4.5 Embarrassment 49114.1.4.6 Pride 49114.1.4.7 Indignation and Anger 49114.1.4.8 Contempt 49114.1.4.9 Pity and Compassion 49214.1.4.10 Awe and Elevation 49214.1.4.11 Gratitude 49214.1.5 Emotion-Provoking Stimuli 49214.2 Effect of Emotion on the Brain 49414.2.1 ERP Change Due to Emotion 49414.2.2 Changes of Normal Brain Rhythms with Emotion 49714.2.3 Emotion and Lateral Brain Engagement 49814.2.4 Perception of Odours and Emotion: Why Are They Related? 49814.3 Emotion-Related Brain Signal Processing and Machine Learning 49914.3.1 Evaluation of Emotion Based on the Changes in Brain Rhythms 50014.3.2 Brain Asymmetricity and Connectivity for Emotion Evaluation 50114.3.3 Changes in ERPs for Emotion Recognition 50414.3.4 Combined Features for Emotion Analysis 50414.4 Other Physiological Measurement Modalities Used for Emotion Study 50714.5 Applications 51014.6 Pain Assessment Using EEG 51014.7 Emotion Elicitation and Induction through Virtual Reality 51214.8 Summary 513References 51415 EEG Analysis of Neurodegenerative Diseases 52515.1 Introduction 52515.2 Alzheimer's Disease 52715.2.1 Application of Brain Connectivity Estimation to AD and MCI 52815.2.2 ERP-Based AD Monitoring 53215.2.3 Other Approaches to EEG-Based AD Monitoring 53215.3 Motor Neuron Disease 53715.4 Parkinson's Disease 53715.5 Huntington's Disease 54115.6 Prion Disease 54215.7 Behaviour Variant Frontotemporal Dementia 54415.8 Lewy Body Dementia 54515.9 Summary 545References 54616 EEG As A Biomarker for Psychiatric and Neurodevelopmental Disorders 55116.1 Introduction 55116.1.1 History 55116.1.1.1 Different Psychiatric and Neurodevelopmental Disorders 55316.1.1.2 NDD Diagnosis 55416.2 EEG Analysis for Different NDDs 55416.2.1 ADHD 55416.2.1.1 ADHD Symptoms and Possible Treatment 55416.2.1.2 EEG-Based Diagnosis of ADHD 55516.2.2 ASD 55916.2.2.1 ASD Symptoms and Possible Treatment 55916.2.2.2 EEG-Based Diagnosis of ASD 56016.2.3 Mood Disorder 56116.2.3.1 EEG for Monitoring Depression 56216.2.3.2 EEG for Monitoring Bipolar Disorder 56416.2.4 Schizophrenia 56516.2.4.1 Schizophrenia Symptoms and Management 56516.2.4.2 EEG as the Biomarker for Schizophrenia 56616.2.5 Anxiety (and Panic) Disorder 56816.2.5.1 Definition and Symptoms 56816.2.5.2 EEG for Assessing Anxiety 56916.2.6 Insomnia 57116.2.6.1 Symptoms of Insomnia 57116.2.6.2 EEG for Insomnia Analysis 57216.2.7 Schizotypal Personality Disorder 57216.2.7.1 What Is Schizotypal Disorder? 57216.2.7.2 EEG Manifestation of Schizotypal 57316.3 Summary 573References 57417 Brain-Computer Interfacing Using EEG 58117.1 Introduction 58117.1.1 State of the Art in BCI 58417.1.2 BCI Terms and Definitions 58517.1.3 Popular BCI Directions 58517.1.4 Virtual Environment for BCI 58617.1.5 Evolution of BCI Design 58717.2 BCI-Related EEG Components 58817.2.1 Readiness Potential and Its Detection 58817.2.2 ERD and ERS 58817.2.3 Transient Beta Activity after the Movement 59317.2.4 Gamma Band Oscillations 59317.2.5 Long Delta Activity 59317.2.6 ERPs 59417.3 Major Problems in BCI 59417.3.1 Preprocessing of the EEGs 59517.4 Multidimensional EEG Decomposition 59717.4.1 Space-Time-Frequency Method 59917.4.2 Parallel Factor Analysis 59917.5 Detection and Separation of ERP Signals 60117.6 Estimation of Cortical Connectivity 60317.7 Application of Common Spatial Patterns 60617.8 Multiclass Brain-Computer Interfacing 60917.9 Cell-Cultured BCI 61017.10 Recent BCI Applications 61017.11 Neurotechnology for BCI 61417.12 Joint EEG and Other Brain-Scanning Modalities for BCI 61717.12.1 Joint EEG-fNIRS for BCI 61717.12.2 Joint EEG-MEG for BCI 61817.13 Performance Measures for BCI Systems 61817.14 Summary 619References 62018 Joint Analysis of EEG and Other Simultaneously Recorded Brain Functional Neuroimaging Modalities 63118.1 Introduction 63118.2 Fundamental Concepts 63118.2.1 Functional Magnetic Resonance Imaging 63118.2.1.1 Blood Oxygenation Level Dependence 63318.2.1.2 Popular fMRI Data Formats 63518.2.1.3 Preprocessing of fMRI Data 63518.2.2 Functional Near-Infrared Spectroscopy 63618.2.3 Magnetoencephalography 64018.3 Joint EEG-fMRI 64018.3.1 Relation Between EEG and fMRI 64018.3.2 Model-Based Method for BOLD Detection 64218.3.3 Simultaneous EEG-fMRI Recording: Artefact Removal from EEG 64418.3.3.1 Gradient Artefact Removal from EEG 64418.3.3.2 Ballistocardiogram Artefact Removal from EEG 64518.3.4 BOLD Detection in fMRI 65218.3.4.1 Implementation of Different NMF Algorithms for BOLD Detection 65318.3.4.2 BOLD Detection Experiments 65418.3.5 Fusion of EEG and fMRI 65918.3.5.1 Extraction of fMRI Time Course from EEG 65918.3.5.2 Fusion of EEG and fMRI; Blind Approach 65918.3.5.3 Fusion of EEG and fMRI; Model-Based Approach 66418.3.6 Application to Seizure Detection 66418.3.7 Investigation of Decision Making in the Brain 66618.3.8 Application to Schizophrenia 66618.3.9 Other Applications 66718.4 EEG-NIRS Joint Recording and Fusion 66818.5 MEG-EEG Fusion 67218.6 Summary 672References 673Index 681

O autorze

Saeid Sanei, PhD, DIC, FBCS, is Professor of Signal Processing and Machine Learning at Nottingham Trent University, UK, and a Visiting Professor at Imperial College London, UK. He received his doctorate in Biomedical Signal and Image Processing from Imperial College London in 1991. He is an internationally renowned expert in signal processing, biomedical signal processing, and pattern recognition.Jonathon A. Chambers, FREng, FIEEE, DSc (Imperial), is Emeritus Professor of Signal and Information Processing within the College of Science and Engineering at the University of Leicester, UK. His research interests are focused upon adaptive signal processing and machine learning and their application in biomedicine, communications, defense, and navigation systems.