Brain-Computer Interfaces

1st Edition - November 5, 2024
Editors: Ayman S. El-Baz, Jasjit S. Suri
Language: English
Paperback ISBN:
9 7 8 - 0 - 3 2 3 - 9 5 4 3 9 - 6
eBook ISBN:
9 7 8 - 0 - 3 2 3 - 9 5 4 4 0 - 2

Advances in Neural Engineering: Brain-Computer Interfaces, Volume Two covers the broad spectrum of neural engineering subfields and applications. The set provides a comprehen… Read more

Purchase options

Institutional subscription on ScienceDirect

Request a sales quote

Advances in Neural Engineering: Brain-Computer Interfaces, Volume Two covers the broad spectrum of neural engineering subfields and applications. The set provides a comprehensive review of dominant feature extraction methods and classification algorithms in the brain-computer interfaces for motor imagery tasks. The book's authors discuss existing challenges in the domain of motor imagery brain-computer interface and suggest possible research directions. The field of neural engineering deals with many aspects of basic and clinical problems associated with neural dysfunction, including sensory and motor information, stimulation of the neuromuscular system to control muscle activation and movement, analysis and visualization of complex neural systems, and more.

Title of Book
Cover image
Title page
Table of Contents
Front matter
Titles in this Series
Series Editors
Copyright
Contributors
Chapter 1. Advances in human activity recognition: Harnessing machine learning and deep learning with topological data analysis
1 Introduction
2 Literature review
3 Background
3.1 Imbalanced data and oversampling techniques
3.2 Features engineering and dimensionality reduction techniques
3.2.1 Topological data analysis (TDA)
3.3 Feature scaling techniques
3.4 Classification, optimization, and performance evaluation
3.4.1 Machine learning classifiers
3.4.2 Deep learning classifiers
3.4.3 K-fold cross-validation
3.4.4 Grid search hyperparameter optimization
4 Methodology
4.1 Data acquisition phase
4.2 Data pre-processing phase
4.2.1 Data balancing
4.2.2 Data sampling
4.3 Features engineering and features extraction using TDA
4.4 ML classification and optimization phase
4.4.1 Hyperparameters optimization using grid search (GS)
4.4.2 Features scaling
4.4.3 Performance improvement
4.5 DL classification phase
5 Experiments and discussions
5.1 First category experiments
5.2 Second category experiments
6 Conclusion, limitation, and future work
Chapter 2. Design and validation of a hybrid programmable platform for the acquisition of ExG signals (EEG, ECG and EMG)
1 Introduction
2 Architecture
2.1 System specifications
2.2 System integration
2.3 Power supply
2.4 First common gain stage
2.5 Filtering stage
2.6 EEG, ECG, and EMG conditioning path
2.7 Noise analysis
2.8 Choice of ADC
2.9 SNR and ENOB calculation
3 Results
3.1 Gain stage
3.2 Time domain analysis
3.3 Evaluation of THD, SFDR, SINAD, and ENOB
3.4 Bode diagrams: AC sweep analysis
3.5 General performance
3.6 Noise analysis
4 Conclusions
Chapter 3. FBSE-based automated classification of motor imagery EEG signals in brain–computer interface
1 Introduction
1.1 Clinical approach for MI-EEG classification
2 Proposed framework
2.1 Database used and pre-processing
2.1.1 Database used
2.1.2 Pre-processing
2.2 Automated MI-EEG classification model
2.2.1 Fourier-Bessel series expansion method
2.2.2 Boundary detection and rhythm separation
2.2.3 Feature extraction and selection
2.3 Classification
2.3.1 Random forest classifier
2.3.2 Linear discriminant analysis classifier
2.3.3 Support vector machine classifier
2.3.4 K-nearest neighbors classifier
2.3.5 Ensemble KNN classifier
2.4 Classification evaluation indexes
2.5 K-fold cross-validation
3 Results
3.1 Overall classification performance
4 Discussions
5 Conclusions
Chapter 4. Automated detection of brain disease using quantum machine learning
1 Introduction
2 Neurodegenerative diseases
2.1 Alzheimer's disease
3 Deep neural network
3.1 Categories of machine learning
3.2 Classification in deep neural network
3.3 The learning procedure
3.4 Convolutional neural networks
4 Quantum computing nurturing artificial intelligence
4.1 Quantum machine learning
4.2 Transfer learning
4.3 Variational quantum circuit
4.4 Dressed quantum circuit
5 Results
5.1 Experiment no. 1: Alzheimer's detection using a pretrained network
5.1.1 GoogleNet model
5.1.2 ResNet34
5.2 Experiment no. 2: Quantum neural network
6 Conclusion
Chapter 5. A study of the relationship of wavelet transform parameters and their impact on EEG classification performance
1 Introduction
2 Background
2.1 Motor imagery
2.2 Time series EEG recording
2.3 Frequency representation
2.4 Time-frequency representation
2.5 Wavelet transform
2.6 CNN-based classification
2.7 Dataset
3 Methodology
3.1 CWT architecture
3.2 Channel selection
3.3 Mother wavelet and scale selection
3.4 Time window length selection
4 Results and discussion
4.1 Channel selection
4.2 Mother wavelet and scale selection
4.3 Time window length selection
5 Conclusions
Chapter 6. BCIs for stroke rehabilitation
1 Introduction
1.1 Stroke
1.2 Brain–computer interfaces
2 BCI-FES therapy
2.1 How does it work?
2.1.1 Therapy protocol
2.1.2 Signal Processing
2.2 Does it work?
3 Results
3.1 Early beginnings
3.2 BCI study
3.3 Application of the BCI treatment
4 Discussion and outlook
Chapter 7. Decoding imagined speech for EEG-based BCI
1 Introduction
1.1 What brain–computer interfaces are?
1.2 Early genealogy of brain–computer interfaces
2 Background on brain–computer interfaces
3 BCI, state of the art
4 Main applications
5 Description of an EEG-based imagined speech BCI
5.1 Basic components
5.2 Feature extraction methods
5.3 Recognition and classification of signals
5.4 Application example
6 Typical imagined speech problems
7 Study cases
7.1 Imagined speech recognition focused on isolated words
7.2 Selecting channels to improve the process
7.3 Incremental learning for imagined speech
7.4 Recognizing words on the continuous wave
Chapter 8. A comparison of deep learning methods and conventional methods for classification of SSVEP signals in brain-computer interface framework
1 Introduction
2 Methodology
2.1 BCI dataset and pre-processing
2.2 Traditional classifiers
2.3 Deep learning methods
2.4 Approach
3 Results
4 Discussion and Conclusion
Chapter 9. Benchmarking convolutional neural networks on continuous EEG signals: The case of motor imagery–based BCI
1 Introduction
2 Methodology
2.1 Dataset
2.2 Preprocessing
2.3 Signal transformation
2.4 CNN models comparison
3 Results
3.1 Performance analysis
3.2 Comparative study
4 Discussion
5 Conclusion
Chapter 10. Advancements in the diagnosis of Alzheimer’s disease (AD) through biomarker detection
1 Introduction
1.1 Overview of AD biomarkers
1.2 Current AD diagnostic techniques
2 Positron emission tomography for AD biomarker detection
2.1 PET scan clinical process
2.2 AD biomarker PET tracers
2.3 Current research in AD diagnosis using PET scans
3 Magnetic resonance imaging (MRI)
3.1 MRI scan clinical process
3.2 AD biomarker MRI tracers
3.3 Current research in AD diagnosis using MRI scans
4 Biosensors for AD detection
4.1 Current AD biomarker biosensor research
4.1.1 Electrochemical biosensors for AD biomarker detection
4.1.2 Optical biosensors for AD biomarker detection
4.2 Future work and clinical outlook for biosensor AD diagnosis
5 Conclusion
Chapter 11. Alcoholism identification by processing the EEG signal using oscillatory modes decomposition and machine learning
1 Introduction
2 Methods and material
2.1 Dataset
2.2 Empirical mode decomposition
2.3 Variational mode decomposition
2.4 Overview of SODP
2.5 Feature space designing approach
2.6 Classification
2.6.1 Support vector machine
2.6.2 Multilayer perceptron neural network
2.6.3 K-nearest neighbor
2.6.4 Random forest
2.6.5 Evaluation measures
3 Results and discussion
4 Conclusion
Chapter 12. Investigating the role of cortical rhythms in modulating kinematic synergies and exploring their potential for stroke rehabilitation
1 Introduction
2 Methodologies
2.1 Experimental protocol
2.2 Derivation of synergies
2.3 Neural features extraction
2.4 Neural decoding
3 Results
4 Discussion and conclusion
Chapter 13. Stimulus-independent noninvasive BCI based on EEG patterns of inner speech
1 Introduction
2 Experimental results
2.1 EEG recording
2.2 Research of the spoken and inner speech-related EEG connectivity in different spatial direction
2.2.1 Experimental procedure
2.2.2 EEG analysis
2.2.3 Methods for classifying EEG coherence patterns
2.2.4 Results of the spoken and inner speech-related EEG connectivity in different spatial direction study
2.2.5 Recognition and classification of EEG coherence patterns
2.3 Studying EEG coherence during real and imagined pronunciation of words and pseudowords
2.3.1 Experimental procedure
2.3.2 EEG analysis
2.3.3 Classification methods for EEG coherence patterns
2.3.4 The results of the studying EEG coherence in real and imagined pronunciation of words and pseudowords
2.3.5 Classification of EEG coherence patterns
3 Conclusions
Chapter 14. A review on contemporary brain–computer interface researches and limitations
1 Introduction
2 Features of brain–computer interface
3 Brain–computer interface working nature
4 Discussion
5 Conclusion
Chapter 15. Noninvasive brain–computer interfaces using fNIRS, EEG, and hybrid EEG-fNIRS
1 Introduction
2 BCI using fNIRS
2.1 Data acquisition
2.2 Preprocessing
2.3 Feature extraction
2.4 Classification
2.5 fNIRS-BCI applications
2.5.1 Motor restoration
2.5.2 Neurorehabilitation
2.5.3 Communication
2.5.4 Neuroergonomics
2.5.5 Other applications
2.6 Challenges and future directions
3 BCI using EEG
3.1 Data acquisition
3.1.1 Slow cortical potential
3.1.2 Sensorimotor rhythms
3.1.3 Evoked potentials
3.2 Preprocessing
3.3 Feature extraction
3.4 Classification
3.5 EEG-BCI applications
3.5.1 Neurorobotics
3.5.2 Cursor control
3.5.3 EEG speller
3.5.4 Emotion recognition
3.5.5 EEG-based gaming
3.6 Challenges and future directions
4 BCI using hybrid EEG-fNIRS
4.1 Data acquisition
4.1.1 EEG signal
4.1.2 fNIRS signal
4.1.3 Fusion of EEG and fNIRS
4.1.4 EEG-informed fNIRS analysis
4.1.5 fNIRS-informed EEG source imaging analysis
4.1.6 Parallel analysis of EEG-fNIRS
4.2 Preprocessing
4.3 Feature extraction
4.4 Classification
4.5 hBCI applications
4.5.1 Gait and balance control
4.5.2 Parkinson's disease
4.5.3 Rehabilitation
4.6 Challenges and future direction
Chapter 16. EEG-based cognitive fatigue recognition using relevant multi-domain features and machine learning
1 Introduction
2 Methods and materials
2.1 Database and visual marking of CF
2.2 Main parts of cognitive fatigue pattern recognition using machine learning techniques
2.2.1 Preprocessing step
2.2.2 Features extraction
2.2.3 Training-testing splitting with cross-validation
2.3 Machines learning techniques
2.3.1 Support vector machines
2.3.2 Multilayer perceptron
2.3.3 Gaussian Naive Bayes
2.4 Performance evaluation of ML models
3 Results
4 Discussion
5 Conclusion and perspectives
Chapter 17. Passive brain–computer interfaces for cognitive and pathological brain physiological states monitoring and control
1 Introduction
2 Basic principles of passive BCIs
3 Passive BCI for monitoring the educational process
3.1 Methodology of the biomarker calculation
3.2 Examples of biomarker implementation for monitoring attention stability
4 Monitoring and controlling attention using pBCI
4.1 Design of the experiment
4.2 EEG analysis procedure
4.3 Results of EEG analysis
4.4 The principle of pBCI for monitoring and controlling attention
5 Passive BCI for depression monitoring
5.1 Biopotentials
5.2 From unimodal to multimodal data collecting systems
5.3 Evoked potentials and multimodal method for emotion classification
6 Conclusion
Chapter 18. Beyond brainwaves: Recommendations for integrating robotics and virtual reality for EEG-driven brain–computer interfaces
1 Introduction
2 Theoretical and design considerations
2.1 Theoretical framework
2.2 Case study: Integrating BCI, VR, and robotics
2.2.1 The visual stimuli
2.2.2 The kinesthetic stimuli
2.2.3 The case for BCI-VR-robot integration
2.3 Networking and data processing pipeline
3 A case-study for EEG-BCI virtual reality integration
4 A case study for EEG–BCI robot integration
5 Ten (10) recommendations for future BCI researchers
5.1 Multimodal engagement and modular flexibility
5.2 Risk and safety evaluations
5.3 Task design and control resolution
5.4 Data processing and machine learning algorithms
5.5 EEG–BCI artifacts attributed to VR and robotics
5.6 Use of dry or standard (gel) EEG electrodes
5.7 Potential effects of enhanced visual and kinesthetic stimuli
5.8 Factors affecting BCI training and performance
5.9 Training task design
5.10 Clinical perspectives
6 Conclusion
Chapter 19. A sociotechnical systems’ perspective to support brain–computer interface development
1 Introduction
2 Introduction to systems thinking
2.1 Sociotechnical system design principles
3 Analyzing the BCI system lifecycle
4 Sociotechnical system-wide risks of BCIs
5 Toward a model of BCI risk sources
6 BCI design insights to enhance safety and performance
6.1 Allocation of tasks between humans and BCIs
6.2 Avoiding the fragmentation of tasks and supporting stakeholder coordination and flows of information
6.3 Controlling the source of problems
6.4 Supporting adaptability through multifunctionality of BCI system elements
6.5 Supporting adaptability though flexible means of undertaking specified tasks by responsible stakeholders
6.6 Alignment of congruent system elements with overall goals
7 Conclusion
Chapter 20. Assessing systemic benefit and risk in the development of BCI neurotechnology
1 BCI innovation and ethics on a global stage
2 BCI beyond the brain
3 BCI as social and systemic
4 Systemic implementations and their implications
5 Systematic ethics
6 A framework for risk management
7 Implementing neurotechnology guidelines
Chapter 21. Recent development of single-channel EEG-based automated sleep stage classification: Review and future perspectives
1 Introduction
2 Concerns regarding human-expert-based manual SLEEP staging
2.1 Investigations of the state-of-the-art models
2.2 Category 1: Time and Frequency domain–based feature extraction with conventional ML techniques
2.3 Category 2: RNN-based models (LSTM, GRU) without attention mechanism
2.4 Category 3: CNN-based feature extraction without attention mechanism
2.5 Category 4: CNN and RNN combined model architecture
2.6 Category 5: models with attention mechanism
2.7 Category 6: Use of Transfer Learning
3 Discussions
3.1 Necessity to train on small dataset
3.2 Development of personalized model
3.3 Necessity of interpretable model
3.4 Requirement for long-term monitoring
3.5 Generalization of source and bias reduction
3.6 Quantification of uncertainty and recursive upgradation of model
4 Conclusion
Index

Ayman S. El-Baz

Dr. El-Baz is a Professor, University Scholar, and Chair of the Bioengineering Department at the University of Louisville, KY. Dr. El-Baz earned his bachelor's and master’s degrees in Electrical Engineering in 1997 and 2001, respectively. He earned his doctoral degree in electrical engineering from the University of Louisville in 2006. In 2009, Dr. El-Baz was named a Coulter Fellow for his contributions to the field of biomedical translational research. Dr. El-Baz has 15 years of hands-on experience in the fields of bio-imaging modeling and non-invasive computer-assisted diagnosis systems. He has authored or coauthored more than 450 technical articles (105 journals, 15 books, 50 book chapters, 175 refereed-conference papers, 100 abstracts, and 15 US patents).

Affiliations and expertise

University of Louisville, USA

Jasjit S. Suri

Dr. Jasjit Suri, PhD, MBA, is an innovator, visionary, scientist, and internationally known world leader. Dr Suri received the Director General’s Gold medal in 1980 and Fellow of (i) American Institute of Medical and Biological Engineering, awarded by the National Academy of Sciences, Washington DC, (ii) Institute of Electrical and Electronics Engineers, (iii) American Institute of Ultrasound in Medicine, (iv) Society of Vascular Medicine, (v) Asia Pacific Vascular Society, and (vi) Asia Association of Artificial Intelligence. Dr. Suri was honored with life time achievement awards by Marcus, NJ, USA and Graphics Era University, Dehradun, India. He has published nearly 300 peer-reviewed Artificial Intelligence articles, nearly 2000 Google Scholar Publications, 100 books, and 100 innovations/trademarks leading to an H-index of nearly 100 with about 43,000 citations. He has held positions as chairman of AtheroPoint, CA, USA, IEEE Denver section, Colorado, USA, and advisory board member to healthcare industries and several universities in the United States of America and abroad.

Affiliations and expertise

Chairman, AtheroPoint LLC, USA

Life Sciences

Physical Sciences & Engineering

Social Sciences & Humanities

Health

Brain-Computer Interfaces

Purchase options

Institutional subscription on ScienceDirect

Ayman S. El-Baz

Jasjit S. Suri