BRAIN-COMPUTER INTERFACE : using deep learning applications

BRAIN-COMPUTER INTERFACE It covers all the research prospects and recent advancements in the brain-computer interface using deep learning. The brain-computer interface (BCI) is an emerging technology that is developing to be more functional in practice. The aim is to establish, through experiences with electronic devices, a communication channel bridging the human neural networks within the brain to the external world. For example, creating communication or control applications for locked-in patients who have no control over their bodies will be one such use. Recently, from communication to marketing, recovery, care, mental state monitoring, and entertainment, the possible application areas have been expanding. Machine learning algorithms have advanced BCI technology in the last few decades, and in the sense of classification accuracy, performance standards have been greatly improved. For BCI to be effective in the real world, however, some problems remain to be solved. Research focusing on deep learning is anticipated to bring solutions in this regard. Deep learning has been applied in various fields such as computer vision and natural language processing, along with BCI growth, outperforming conventional approaches to machine learning. As a result, a significant number of researchers have shown interest in deep learning in engineering, technology, and other industries; convolutional neural network (CNN), recurrent neural network (RNN), and generative adversarial network (GAN). Audience Researchers and industrialists working in brain-computer interface, deep learning, machine learning, medical image processing, data scientists and analysts, machine learning engineers, electrical engineering, and information technologists. Cover 1 Title Page 5 Copyright Page 6 Contents 7 Preface 15 Chapter 1 Introduction to Brain–Computer Interface: Applications and Challenges 17 1.1 Introduction 17 1.2 The Brain – Its Functions 19 1.3 BCI Technology 19 1.3.1 Signal Acquisition 21 1.3.1.1 Invasive Methods 22 1.3.1.2 Non-Invasive Methods 24 1.3.2 Feature Extraction 26 1.3.3 Classification 27 1.3.3.1 Types of Classifiers 28 1.4 Applications of BCI 29 1.5 Challenges Faced During Implementation of BCI 33 References 37 Chapter 2 Introduction: Brain–Computer Interface and Deep Learning 41 2.1 Introduction 42 2.1.1 Current Stance of P300 BCI 44 2.2 Brain–Computer Interface Cycle 45 2.3 Classification of Techniques Used for Brain–Computer Interface 54 2.3.1 Application in Mental Health 54 2.3.2 Application in Motor-Imagery 54 2.3.3 Application in Sleep Analysis 55 2.3.4 Application in Emotion Analysis 55 2.3.5 Hybrid Methodologies 56 2.3.6 Recent Notable Advancements 57 2.4 Case Study: A Hybrid EEG-fNIRS BCI 62 2.5 Conclusion, Open Issues and Future Endeavors 63 References 65 Chapter 3 Statistical Learning for Brain–Computer Interface 79 3.1 Introduction 80 3.1.1 Various Techniques to BCI 80 3.1.1.1 Non-Invasive 80 3.1.1.2 Semi-Invasive 81 3.1.1.3 Invasive 83 3.2 Machine Learning Techniques to BCI 83 3.2.1 Support Vector Machine (SVM) 85 3.2.2 Neural Networks 85 3.3 Deep Learning Techniques Used in BCI 86 3.3.1 Convolutional Neural Network Model (CNN) 88 3.3.2 Generative DL Models 89 3.4 Future Direction 89 3.5 Conclusion 90 References 91 Chapter 4 The Impact of Brain–Computer Interface on Lifestyle of Elderly People 93 4.1 Introduction 94 4.2 Diagnosing Diseases 95 4.3 Movement Control 100 4.4 IoT 101 4.5 Cognitive Science 102 4.6 Olfactory System 104 4.7 Brain-to-Brain (B2B) Communication Systems 105 4.8 Hearing 106 4.9 Diabetes 107 4.10 Urinary Incontinence 108 4.11 Conclusion 109 References 109 Chapter 5 A Review of Innovation to Human Augmentation in Brain-Machine Interface – Potential, Limitation, and Incorporation of AI 117 5.1 Introduction 118 5.2 Technologies in Neuroscience for Recording and Influencing Brain Activity 119 5.2.1 Brain Activity Recording Technologies 120 5.2.1.1 A Non-Invasive Recording Methodology 120 5.2.1.2 An Invasive Recording Methodology 120 5.3 Neuroscience Technology Applications for Human Augmentation 122 5.3.1 Need for BMI 122 5.3.1.1 Need of BMI Individuals for Re-Establishing the Control and Communication of Motor 123 5.3.1.2 Brain-Computer Interface Noninvasive Research at Wadsworth Center 123 5.3.1.3 An Interface of Berlin Brain-Computer: Machine Learning-Dependent of User-Specific Brain States Detection 123 5.4 History of BMI 124 5.5 BMI Interpretation of Machine Learning Integration 127 5.6 Beyond Current Existing Methodologies: Nanomachine Learning BMI Supported 132 5.7 Challenges and Open Issues 135 5.8 Conclusion 136 References 137 Chapter 6 Resting-State fMRI: Large Data Analysis in Neuroimaging 143 6.1 Introduction 144 6.1.1 Principles of Functional Magnetic Resonance Imaging (fMRI) 144 6.1.2 Resting State fMRI (rsfMRI) for Neuroimaging 144 6.1.3 The Measurement of Fully Connected and Construction of Default Mode Network (DMN) 145 6.2 Brain Connectivity 145 6.2.1 Anatomical Connectivity 145 6.2.2 Functional Connectivity 146 6.3 Better Image Availability 146 6.3.1 Large Data Analysis in Neuroimaging 147 6.3.2 Big Data rfMRI Challenges 149 6.3.3 Large rfMRI Data Software Packages 150 6.4 Informatics Infrastructure and Analytical Analysis 153 6.5 Need of Resting-State MRI 153 6.5.1 Cerebral Energetics 153 6.5.2 Signal to Noise Ratio (SNR) 153 6.5.3 Multi-Purpose Data Sets 154 6.5.4 Expanded Patient Populations 154 6.5.5 Reliability 154 6.6 Technical Development 154 6.7 rsfMRI Clinical Applications 155 6.7.1 Mild Cognitive Impairment (MCI) and Alzheimer’s Disease (AD) 155 6.7.2 Fronto-Temporal Dementia (FTD) 156 6.7.3 Multiple Sclerosis (MS) 157 6.7.4 Amyotrophic Lateral Sclerosis (ALS) and Depression 159 6.7.5 Bipolar 160 6.7.6 Schizophrenia 161 6.7.7 Attention Deficit Hyperactivity Disorder (ADHD) 163 6.7.8 Multiple System Atrophy (MSA) 163 6.7.9 Epilepsy/Seizures 163 6.7.10 Pediatric Applications 165 6.8 Resting-State Functional Imaging of Neonatal Brain Image 165 6.9 Different Groups in Brain Disease 167 6.10 Learning Algorithms for Analyzing rsfMRI 167 6.11 Conclusion and Future Directions 170 References 170 Chapter 7 Early Prediction of Epileptic Seizure Using Deep Learning Algorithm 173 7.1 Introduction 174 7.2 Methodology 180 7.3 Experimental Results 185 7.4 Taking Care of Children with Seizure Disorders 188 7.5 Ketogenic Diet 188 7.6 Vagus Nerve Stimulation (VNS) 188 7.7 Brain Surgeries 189 7.8 Conclusion 189 References 191 Chapter 8 Brain–Computer Interface-Based Real-Time Movement of Upper Limb Prostheses Topic: Improving the Quality of the Elderly with Brain-Computer Interface 195 8.1 Introduction 196 8.1.1 Motor Imagery Signal Decoding 197 8.2 Literature Survey 198 8.3 Methodology of Proposed Work 200 8.3.1 Proposed Control Scheme 201 8.3.2 One Versus All Adaptive Neural Type-2 Fuzzy Inference System (OVAANT2FIS) 203 8.3.3 Position Control of Robot Arm Using Hybrid BCI for Rehabilitation Purpose 203 8.3.4 Jaco Robot Arm 205 8.3.5 Scheme 1: Random Order Positional Control 205 8.4 Experiments and Data Processing 208 8.4.1 Feature Extraction 211 8.4.2 Performance Analysis of the Detectors 213 8.4.3 Performance of the Real Time Robot Arm Controllers 214 8.5 Discussion 216 8.6 Conclusion and Future Research Directions 218 References 219 Chapter 9 Brain–Computer Interface-Assisted Automated Wheelchair Control Management-Cerebro: A BCI Application 221 9.1 Introduction 222 9.1.1 What is a BCI? 223 9.2 How Do BCI’s Work? 223 9.2.1 Measuring Brain Activity 224 9.2.1.1 Without Surgery 224 9.2.1.2 With Surgery 224 9.2.2 Mental Strategies 225 9.2.2.1 SSVEP 226 9.2.2.2 Neural Motor Imagery 226 9.3 Data Collection 227 9.3.1 Overview of the Data 227 9.3.2 EEG Headset 229 9.3.3 EEG Signal Collection 230 9.4 Data Pre-Processing 231 9.4.1 Artifact Removal 232 9.4.2 Signal Processing and Dimensionality Reduction 233 9.4.3 Feature Extraction 233 9.5 Classification 234 9.5.1 Deep Learning (DL) Model Pipeline 235 9.5.2 Architecture of the DL Model 236 9.5.3 Output Metrics of the Classifier 237 9.5.4 Deployment of DL Model 237 9.5.5 Control System 239 9.5.6 Control Flow Overview 239 9.6 Control Modes 239 9.6.1 Speech Mode 239 9.6.2 Blink Stimulus Mapping 239 9.6.3 Text Interface 241 9.6.4 Motion Mode 241 9.6.5 Motor Arrangement 241 9.6.6 Imagined Motion Mapping 242 9.7 Compilation of All Systems 242 9.8 Conclusion 242 References 243 Chapter 10 Identification of Imagined Bengali Vowels from EEG Signals Using Activity Map and Convolutional Neural Network 247 10.1 Introduction 248 10.1.1 Electroencephalography (EEG) 249 10.1.2 Imagined Speech or Silent Speech 249 10.2 Literature Survey 250 10.3 Theoretical Background 254 10.3.1 Convolutional Neural Network 254 10.3.2 Activity Map 256 10.4 Methodology 258 10.4.1 Data Collection 259 10.4.2 Pre-Processing 260 10.4.3 Feature Extraction 261 10.4.4 Classification 263 10.5 Results 265 10.6 Conclusion 268 Acknowledgment 268 References 268 Chapter 11 Optimized Feature Selection Techniques for Classifying Electrocorticography Signals 271 11.1 Introduction 272 11.1.1 Brain–Computer Interface 272 11.2 Literature Study 274 11.3 Proposed Methodology 276 11.3.1 Dataset 277 11.3.2 Feature Extraction Using Auto-Regressive (AR) Model and Wavelet Transform 277 11.3.2.1 Auto-Regressive Features 277 11.3.2.2 Wavelet Features 278 11.3.2.3 Feature Selection Methods 278 11.3.2.4 Information Gain (IG) 279 11.3.2.5 Clonal Selection 279 11.3.2.6 An Overview of the Steps of the CLONALG 280 11.3.3 Hybrid CLONALG 281 11.4 Experimental Results 284 11.4.1 Results of Feature Selection Using IG with Various Classifiers 288 11.4.2 Results of Optimizing Support Vector Machine Using CLONALG Selection 290 11.5 Conclusion 292 References 293 Chapter 12 BCI – Challenges, Applications, and Advancements 295 12.1 Introduction 295 12.1.1 BCI Structure 296 12.2 Related Works 297 12.3 Applications 298 12.4 Challenges and Advancements 313 12.5 Conclusion 315 References 315 Index 319 EULA 323