Based on the authors’ groundbreaking research, Automated EEG-Based Diagnosis of Neurological Disorders: Inventing the Future of Neurology presents a research ideology, a novel multi-paradigm methodology, and advanced computational models for the automated EEG-based diagnosis of neurological disorders. It is based on the ingenious integration of three different computing technologies and problem-solving paradigms: neural networks, wavelets, and chaos theory. The book also includes three introductory chapters that familiarize readers with these three distinct paradigms. After extensive research and the discovery of relevant mathematical markers, the authors present a methodology for epilepsy diagnosis and seizure detection that offers an exceptional accuracy rate of 96 percent. They examine technology that has the potential to impact and transform neurology practice in a significant way. They also include some preliminary results towards EEG-based diagnosis of Alzheimer’s disease. The methodology presented in the book is especially versatile and can be adapted and applied for the diagnosis of other brain disorders. The senior author is currently extending the new technology to diagnosis of ADHD and autism. A second contribution made by the book is its presentation and advancement of Spiking Neural Networks as the seminal foundation of a more realistic and plausible third generation neural network.
Author(s): Hojjat Adeli, Samanwoy Ghosh-Dastidar
Edition: 1
Year: 2010
Language: English
Pages: 423
Tags: Медицинские дисциплины;Клинические методы диагностики, лабораторная диагностика;Функциональная диагностика;Электроэнцефалография;
Cover......Page 1
Title Page......Page 3
ISBN 9781439815311......Page 4
Preface......Page 6
Acknowledgments......Page 9
About the Authors......Page 10
List of Figures......Page 12
List of Tables......Page 21
Contents......Page 24
I. Basic Concepts......Page 32
1. Introduction......Page 34
2.1 Signal Digitization and Sampling Rate......Page 36
2.2 Time and Frequency Domain Analyses......Page 38
2.3.1 Short Time Fourier Transform (STFT)......Page 42
2.3.2 Wavelet Transform......Page 43
2.4 Types of Wavelets......Page 51
2.5 Advantages of the Wavelet Transform......Page 57
3.1 Introduction......Page 60
3.2 Attractors in Chaotic Systems......Page 62
3.3.1 Measures of Chaos......Page 71
3.3.2 Preliminary Chaos Analysis - Lagged Phase Space......Page 72
3.3.3 Final Chaos Analysis......Page 76
4.1 Data Classification......Page 80
4.2 Cluster Analysis......Page 81
4.3 k-Means Clustering......Page 85
4.4 Discriminant Analysis......Page 86
4.5 Principal Component Analysis......Page 87
4.6 Artificial Neural Networks......Page 89
4.6.1 Feed forward Neural Network and Error Backpropagation......Page 90
4.6.2 Radial Basis Function Neural Network......Page 97
II. Automated EEG-Based Diagnosis of Epilepsy......Page 100
5.1 Spatio-Temporal Activity in the Human Brain......Page 102
5.2 EEG: A Spatio-Temporal Data Mine......Page 103
5.3 Data Mining Techniques......Page 109
5.4.1 Feature Space Identification and Feature EnhancementUsing Wavelet-Chaos Methodology......Page 112
5.4.2 Development of Accurate and Robust Classifiers......Page 115
5.5 Epilepsy and Epileptic Seizures......Page 116
6.1 Introduction......Page 120
6.2 Wavelet Analysis of a Normal EEG......Page 124
6.3.1 Daubechies Wavelets......Page 127
6.3.2 Harmonic Wavelet......Page 128
6.3.3 Characterization......Page 137
6.4 Concluding Remarks......Page 147
7.1 Introduction......Page 150
7.2 Wavelet-Chaos Analysis of EEG Signals......Page 153
7.3.1 Description of the EEG Data Used in the Research......Page 156
7.3.2 Data Preprocessing and Wavelet Decomposition of EEG into Sub-Bands......Page 159
7.3.3 Results of Chaos Analysis for a Sample Set of Unfiltered EEGs......Page 160
7.3.4 Statistical Analysis of Results for All EEGs......Page 164
7.4 Concluding Remarks......Page 169
8.1 Introduction......Page 174
8.2 Wavelet-Chaos Analysis: EEG Sub-Bands and Feature Space Design......Page 175
8.3 Data Analysis......Page 177
8.4.1 k-Means Clustering......Page 179
8.4.2 Discriminant Analysis......Page 181
8.4.3 RBFNN......Page 184
8.4.4 LMBPNN......Page 186
8.5 Mixed-Band Analysis: Wavelet-Chaos-Neural Network......Page 187
8.6 Concluding Remarks......Page 191
9.1 Introduction......Page 194
9.2 Principal Component Analysis for Feature Enhancement......Page 196
9.3 Cosine Radial Basis Function Neural Network: EEG Classification......Page 200
9.4.2 Output Encoding Scheme......Page 203
9.4.3 Comparison of Classifiers......Page 204
9.4.4 Sensitivity to Number of Eigenvectors......Page 207
9.4.5 Sensitivity to Training Size......Page 208
9.4.6 Sensitivity to Spread......Page 209
9.5 Concluding Remarks and Clinical Significance......Page 212
III. Automated EEG-Based Diagnosis of Alzheimer's Disease......Page 214
10.1 Introduction......Page 216
10.2 Neurological Markers of Alzheimer's Disease......Page 218
10.3.1 Anatomical Imaging versus Functional Imaging......Page 222
10.3.2 Identification of Region of Interest (ROI)......Page 224
10.3.3 Image Registration Techniques......Page 225
10.3.4 Linear and Area Measures......Page 226
10.3.5 Volumetric Measures......Page 227
10.4 Classification Models......Page 228
10.5.1 Approaches to Neural Modeling......Page 231
10.5.2 Hippocampal Models of Associative Memory......Page 233
10.5.3 Neural Models of Progression of AD......Page 234
11.1 EEGs for Diagnosis and Detection of Alzheimer's Disease......Page 238
11.2 Time-Frequency Analysis......Page 239
11.3 Wavelet Analysis......Page 243
11.4 Chace Analysis......Page 244
11.5 Concluding Remarks......Page 249
12.1 Introduction......Page 252
12.2.1 Description of the EEG Data......Page 255
12.2.3 Chaos Analysis and ANOVA Design......Page 257
12.3.1 Complexity and Chaoticity of the EEG: Results of the Three-Way Factorial ANOVA......Page 260
12.3.3 Local Complexity and Chaoticity......Page 261
12.4 Discussion......Page 0
12.4.1 Chaoticity versus Complexity......Page 262
12.4.2 Eyes Open versus Eyes Closed......Page 265
12.5 Concluding Remarks......Page 267
IV. Third Generation Neural Networks: Spiking Neural Networks......Page 270
13.1 Introduction......Page 272
13.2 Information Encoding and Evolution of Spiking Neurons......Page 274
13.3 Mechanism of Spike Generation in Biological Neurons......Page 277
13.4 Models of Spiking Neurons......Page 283
13.5 Spiking Neural Networks (SNNs)......Page 285
13.6 Unsupervised Learning......Page 287
13.7 Supervised Learning......Page 289
13.7.1 Feedforward Stage: Computation of Spike Times and Network Error......Page 295
13.7.2 Backpropagation Stage: Learning Algorithms......Page 297
14.1.1 Number of Neurons in Each Layer......Page 302
14.1.3 Initialization of Weights......Page 303
14.1.4 Heuristic Rules for SNN Learning Algorithms......Page 304
14.2.1 Input and Output Encoding......Page 306
14.2.2 SNN Architecture......Page 307
14.2.4 Type of Neuron (Excitatory or Inhibitory)......Page 309
14.2.5 Convergence Results for a Simulation Time of 50 ms......Page 310
14.2.6 Convergence Results for a Simulation Time of 25 ms......Page 316
14.3.1 Input Encoding......Page 319
14.3.2 Output Encoding......Page 322
14.4.3 Convergence Criteria: MSE and Training Accuracy......Page 330
14.3.4 Convergence Criteria: MSE and Training Accuracy......Page 323
14.3.5 Heuristic Rules for Adaptive Simulation Time and SpikeProp Learning Rate......Page 325
14.3.6 Classification Accuracy and Computational Efficiency versus Training Size......Page 326
14.3.7 Summary......Page 328
14.4.1 Input and Output Encoding......Page 329
14.4.4 Classification Accuracy versus Training Size and Number of Input Neurons......Page 331
14.4.5 Classification Accuracy versus Desired Training Accuracy 301......Page 332
14.4.6 Summary......Page 333
14.5 Concluding Remarks......Page 334
15.1 Introduction......Page 336
15.2.1 MuSpiNN Architecture......Page 341
15.2.2 Multi-Spiking Neuron and the Spike Response Model......Page 343
15.3.1 MuSpi NN Error Function......Page 348
15.3.2 Error Backpropagation for Adjusting Synaptic Weights......Page 349
15.3.3 Gradient Computation for Synapses Between a Neuron in the Last Hidden Layer and a Neuron in the Output Layer......Page 350
15.3.4 Gradient Computation for Synapses Between a Neuronin the Input or Hidden Layer and a Neuron in the Hidden Layer......Page 355
16.1 Parameter Selection and Weight Initialization......Page 360
16.2 Heuristic Rules for Multi-Spike Prop......Page 362
16.3 XOR Problem......Page 363
16.4 Fisher Iris Classification Problem......Page 365
16.5 EEG Classification Problem......Page 369
16.6 Discussion and Concluding Remarks......Page 370
17. The Future......Page 378
Bibliography......Page 380
Index......Page 414
Back Page......Page 419
