Murugappan M., Rajamanicka Y. Biomedical Signals Based Computer-Aided Diagnosis for Neurological Disorders

Cham: Springer, 2022. — 295 p.Biomedical signals provide unprecedented insight into abnormal or anomalous neurological conditions. The computer-aided diagnosis (CAD) system plays a key role in detecting neurological abnormalities and improving diagnosis and treatment consistency in medicine. This book covers different aspects of biomedical signals-based systems used in the automatic detection/identification of neurological disorders. Several biomedical signals are introduced and analyzed, including electroencephalogram (EEG), electrocardiogram (ECG), heart rate (HR), magnetoencephalogram (MEG), and electromyogram (EMG). It explains the role of the CAD system in processing biomedical signals and the application to neurological disorder diagnosis. The book provides the basics of biomedical signal processing, optimization methods, and machine learning/deep learning techniques used in designing CAD systems for neurological disorders.AcknowledgementsAbnormal EEG Detection Using Time-Frequency Images and Convolutional Neural NetworkRelated Studies
Materials and Methods
Data Preparation
Time-Frequency Representation-Based Abnormal EEG Detection
Convolutional Neural Networks
Inception-ResNet-V
DenseNet
SeizureNet
Extreme Learning Machine
Training and Validation
Results
DiscussionPhysical Action Categorization Pertaining to Certain Neurological Disorders Using Machine Learning-Based Signal AnalysisRelated Work
Materials and Methods
Acquisition Band
Data Acquisition
Methodology
Pre-Processing and Segmentation
Feature Vector Generation
Time Domain Features
Inter-Channel Statistics
Maximum Similarity Index
Maximum Covariance Index
Power Spectral Density (PSD)
Log Moments of Fourier Spectra (LMF)
Feature Normalization
Classifier
Experimental Results and Discussion
Dataset
Experimental Results
Discussion
ConclusionsA Comparative Study on EEG Features for Neonatal Seizure DetectionMethodology
Database
Pre-processing
Feature Extraction
Feature Ranking and Classification
Results
Discussions
Effect of Number of Channels
Effect of Feature Extraction Methods
Effect of Feature Ranking Methods
Effect of Classifiers
Effect of Performance Metrics
Limitations and Future DirectionsHilbert Huang Transform (HHT) Analysis of Heart Rate Variability (HRV) in Recognition of Emotion in Children with Autism SpectMethodology
Protocol Design
ECG Data Acquisition Using Wearable Sensor
Data Pre-processing
Feature Extraction
Hilbert-Huang Transform (HHT) Algorithm
Feature Classification
Results
Pre-processing
Extraction of Features
Discussion and ConclusionDetection of Tonic-Clonic Seizures Using Scalp EEG of Spectral MomentsMethods
Feature Extraction
Rectangular Window
Hanning Window
Hamming Window
Flattop Window
Spectral Moments
Classification
Results and DiscussionInvestigation of the Brain Activation Pattern of Stroke Patients and Healthy Individuals During Happiness and SadnessMaterials and Methods
Subjects´ Background
EEG Device
Signal Pre-processing
Feature Analysis
Emotion Classification
Results
Average Values of Hjorth Parameters
Distribution of Hjorth Parameters in Different Frequency Sub-bands
Statistical Analysis of Hjorth Parameters Between Happiness and Sadness
Average Distribution over Frontal Regions of Left and Right Hemispheres
Emotion Classification Using KNN Classifier
Discussion
Emotion Impairment Analyzed Using Hjorth Parameters
Region of Emotional Activation in the Brain
The Involvement of Frontal Region in Emotion ProcessingA Novel Parametric Nonstationary Signal Model for EEG Signals and Its Application in Epileptic Seizure DetectionSignal Modeling and Parametric Estimation
Segmentation and AM-FM Demodulation
Phase Smoothing and Instantaneous Frequency (IF) Estimation
Simulation Results
Applications
Conclusion and DiscussionBiomedical Signal Analysis Using Entropy Measures: A Case Study of Motor Imaginary BCI in End Users with DisabilityEntropy
History
Entropy and Information Theory
Mathematical Formulation
Shannon Entropy
Approximate Entropy
Neural Network Entropy
Numerical Results
Channel Selection
Conclusion and Further WorkAutomatic Detection of Epilepsy Using CNN-GRU Hybrid ModelMethodology
Dataset
Scalogram
Architecture
Training the Models
Results
Two-Class Performance
Case : AB-CDE
Case : ABCD-E
Three-Class Performance
Five-Class Performance
DiscussionCatalogic Systematic Literature Review of Hardware-Accelerated Neurodiagnostic SystemsMethodology
Research Questions
Search Strategy
Selection Criteria
Data Categorization
Quality Assessment
Data Extraction
Review Conduction
Metadata Analysis
Data Analysis
C: Acquisition Stage
C: Acquisition Interface
C: Analog-to-Digital Conversion
C: Amplification
C: Data Compression
C: Transmission Protocol
C: Preprocessing Stage
C: Detrending/Filtering
C: Feature Extraction
C: Processing Stage
C: Seizure Detection
C: Emotion/Vigilance Classification
C: Autism/Anomaly Detection
C: Intention/Imagery/Gesture/Speech Recognition
C: Sleep/Hypnosis-Level Classification
Results
RQ: What Hardware Accelerators Exist?
RQ: Which Algorithms Are Implemented?
RQ: Which Technology/Platform Is Preferred?
RQ: Which Stages Are Hardware Accelerated?
RQ: What Are the Advantages and Disadvantages?
RQ: What Are the Metrics for the Assessment?
DiscussionWearable Real-Time Epileptic Seizure Detection and Warning SystemMaterials and Methods
Block Diagram of the System
Wearable Bio-signal Acquisition Subsystem
Electrodermal Activity (EDA) Sensor
Accelerometer (ACC) Module
Wireless Communication over BLE
Intelligent Epilepsy Detection and Alerting Subsystem
Database Description
Analysis
Data Preprocessing
Feature Extraction
Training, Validation, and Testing
Results and Discussion
Performance Evaluation of EDA and ACC Sensor
Reliability of the BLE Transmission
Power Consumption of the Wearable Module
Classification Results for ACC Data Alone
EDA and ACC Data from Empatica and Prototype System
Classification of Seizure Using EDA Alone, ACC Alone, and Fused EDA-ACC
ConclusionsAnalysis of Intramuscular Coherence of Lower Limb Muscle Activities Using Magnitude Squared CoherenceMethods
Magnitude Squared Coherence
Results
Discussion