Epileptic seizure detection in EEG signals using tunable-Q factor wavelet transform and bootstrap aggregating

Area of Science:

• Biomedical Engineering
• Signal Processing
• Neurology

Background:

• Traditional epileptic seizure detection relies on manual EEG analysis, which is time-consuming and prone to errors.
• Manual analysis is impractical for large epilepsy research datasets.
• Automated methods can improve efficiency, accuracy, and aid clinical diagnosis.

Purpose of the Study:

• To develop a novel automated epilepsy seizure detection scheme using EEG signals.
• To introduce the first application of spectral features in the TQWT domain combined with Bagging for seizure identification.
• To evaluate the performance of the proposed automated detection method.

Main Methods:

• EEG signal segments were decomposed into sub-bands using Tunable-Q factor wavelet transform (TQWT).
• Spectral features were extracted from TQWT sub-bands and analyzed for suitability.
• Epileptic seizures were classified using Bootstrap Aggregating (Bagging) with optimized TQWT and Bagging parameters.
• Performance was evaluated on a benchmark EEG database for various classification scenarios.

Main Results:

• The proposed automated method demonstrated superior performance compared to existing state-of-the-art algorithms.
• Key performance metrics including sensitivity, specificity, and accuracy were significantly improved.
• The study established the efficacy of spectral features within the TQWT domain when combined with Bagging.

Conclusions:

• The developed automated seizure detection system can significantly reduce the workload for medical professionals.
• The method offers a faster and more reliable approach to epilepsy diagnosis.
• This automated scheme holds substantial potential to benefit epilepsy research by enabling efficient analysis of large datasets.