EEG Signal Analysis for Diagnosing Neurological Disorders Using Discrete Wavelet Transform and Intelligent...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Electroencephalogram (EEG) signal analysis is crucial for diagnosing neurological brain disorders.
• Existing methods may require separate analyses for different conditions.
• A unified system for diagnosing multiple neurological disorders simultaneously is needed.

Purpose of the Study:

• To develop a single, accurate system for diagnosing neurological brain disorders, specifically epilepsy and autism spectrum disorder (ASD).
• To investigate various EEG feature-extraction and classification techniques for improved diagnostic accuracy.
• To enable simultaneous diagnosis of one or two neurological diseases (two-class and three-class modes).

Main Methods:

• EEG data preprocessing included artifact removal using Independent Components Analysis (ICA) and filtering with an elliptic band-pass filter.
• Feature extraction involved Discrete Wavelet Transform (DWT) to decompose signals into sub-bands (delta, theta, alpha, beta, gamma), followed by statistical methods (logarithmic band power, standard deviation, variance, kurtosis, Shannon entropy).
• Classification was performed using Linear Discriminant Analysis (LDA), Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Artificial Neural Networks (ANNs).

Main Results:

• The combination of DWT with Shannon Entropy (SE) and Logarithmic Band Power (LBP) yielded the highest accuracy across classifiers.
• Support Vector Machine (SVM) achieved an overall classification accuracy approaching 99.9% for the three-class single-channel mode.
• Artificial Neural Networks (ANNs) reached approximately 97% accuracy for the three-class multi-channel mode.

Conclusions:

• The developed system demonstrates high efficacy in diagnosing epilepsy and ASD from EEG signals.
• The proposed methodology, integrating DWT, statistical features (SE, LBP), and advanced classifiers (SVM, ANN), offers a robust approach for multi-class neurological disorder diagnosis.
• This system holds potential for improving the efficiency and accuracy of neurological disorder diagnosis in clinical settings.