Fast Fractional Fourier Transform-Aided Novel Graphical Approach for EEG Alcoholism Detection.

Area of Science:

• Neuroscience
• Medical Diagnostics
• Signal Processing

Background:

• Alcoholism is a severe brain disorder causing cognitive, emotional, and behavioral issues.
• Traditional alcoholism diagnosis (CAGE assessment) is lengthy, error-prone, and biased.
• There is a need for objective and accurate alcoholism detection methods.

Purpose of the Study:

• To develop a novel framework for identifying alcoholism using electroencephalogram (EEG) signals.
• To overcome the limitations of existing diagnostic methods.
• To establish an efficient and reliable system for real-time alcoholism detection.

Main Methods:

• EEG data preprocessing using multiscale principal component analysis to remove artifacts.
• A graphical technique based on fast fractional Fourier transform coefficients to visualize EEG signal complexity.
• Feature extraction, ensembled feature selection, and neural network classifier evaluation.

Main Results:

• The proposed framework achieved 97.5% accuracy, 96.7% sensitivity, and 98.3% specificity using a recurrent neural network (RNN).
• Sixteen selected features demonstrated effective differentiation between regular and alcoholic EEG signals.
• The method successfully visualized chaotic characteristics and complexities in EEG signals.

Conclusions:

• The novel EEG-based framework offers a highly accurate and reliable method for alcoholism identification.
• This approach can significantly improve upon traditional diagnostic tools.
• The developed framework has the potential for real-time application in clinical and commercial settings.