Seizure classification in EEG signals utilizing Hilbert-Huang transform.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Brain functionality analysis relies on brain imaging or signal analysis.
• Abnormal brain activity, like seizures, involves disturbances in neuronal electrochemical activity and synchronous discharges.
• Physicians need reliable methods to differentiate between healthy and seizure electroencephalographic (EEG) signals.

Purpose of the Study:

• To develop and validate a novel classification algorithm for EEG signals.
• To enable physicians to accurately discriminate between normal and seizure-free EEG signals.
• To provide an efficient diagnostic tool for brain functionality abnormalities.

Main Methods:

• EEG signals from accessible databases were analyzed using MATLAB.
• A classification algorithm based on the Hilbert-Huang Transform was employed to extract local amplitude and frequency information.
• The t-test and Euclidean clustering were used for comparing normal and ictal EEG signal characteristics.

Main Results:

• The proposed method achieved high accuracy (94%) and specificity (96%) in classifying EEG signals.
• Compared to Multivariate Empirical Mode Decomposition (80% accuracy), the new method demonstrated superior performance.
• The t-test yielded a P-value < 0.02, supporting the statistical significance of the findings.

Conclusions:

• An original tool for EEG signal processing has been developed, aiding physicians in diagnosing brain abnormalities.
• The system offers benefits including fast diagnosis, high accuracy, sensitivity, specificity, and user-friendliness.
• The tool's low cost and ease of interface further enhance its utility as an efficient diagnostic solution.