Alcoholic EEG signal classification with Correlation Dimension based distance metrics approach and Modified Adaboost...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computer Science

Background:

• Alcoholism significantly impacts brain function, necessitating reliable diagnostic tools.
• Electroencephalography (EEG) signals offer a window into brain activity but are complex, non-linear, and non-stationary.
• Manual interpretation of EEG signals is time-consuming and challenging due to low signal magnitude and variability.

Purpose of the Study:

• To develop a Computer Aided Diagnosis (CAD) system for distinguishing between normal and alcoholic EEG signals.
• To explore the efficacy of Correlation Dimension (CD) for feature extraction from alcoholic EEG signals.
• To evaluate various classifiers for accurate discrimination of EEG patterns.

Main Methods:

• EEG signals were analyzed using Correlation Dimension (CD) for initial clustering and feature extraction.
• Feature selection was performed using correlation distance, city block distance, cosine distance, and Chebyshev distance.
• Classification was conducted using Adaboost.RT, a modified Adaboost.RT with Ridge/Lasso thresholding, Random Forest, Artificial Neural Networks (ANN), Support Vector Machine (SVM), Naïve Bayesian Classifier (NBC), K-means, and K Nearest Neighbor (KNN).

Main Results:

• Correlation distance metrics combined with CD demonstrated effective discrimination between normal and alcoholic EEG signals.
• The proposed Modified Adaboost.RT classifier, incorporating Ridge-based soft thresholding, achieved a high classification accuracy of 98.99%.
• Various machine learning classifiers showed varying degrees of success in classifying EEG signals.

Conclusions:

• The proposed CAD approach, utilizing CD and modified Adaboost.RT, shows significant promise for accurate alcoholism detection via EEG analysis.
• Non-linear features derived from EEG signals are crucial for effective discrimination.
• This methodology offers a more efficient and objective alternative to manual EEG interpretation for diagnosing alcoholism.