A high-performance seizure detection algorithm based on Discrete Wavelet Transform (DWT) and EEG.

Area of Science:

• Signal Processing
• Biomedical Engineering
• Neurology

Background:

• Discrete Wavelet Transform (DWT) is crucial for analyzing epileptic electroencephalography (EEG) signals, particularly for seizure detection.
• Previous DWT applications in EEG analysis often rely on empirical or arbitrary parameter settings, limiting accuracy and efficiency.
• Optimizing DWT parameters is essential for enhancing the performance of computer-aided seizure detection systems.

Purpose of the Study:

• To develop an automated framework for optimizing Discrete Wavelet Transform (DWT) settings in epileptic electroencephalography (EEG) seizure detection.
• To improve the accuracy and reduce the computational cost of EEG-based seizure detection.
• To establish a systematic approach for selecting optimal wavelet families, decomposition levels, and frequency bands.

Main Methods:

• Decomposition of EEG data using 7 common wavelet families to their maximum theoretical levels.
• Exhaustive search for optimal wavelets, decomposition levels, and frequency bands to maximize accuracy and minimize computational cost.
• Feature selection and redundancy reduction, removing approximately 40% of data.

Main Results:

• The developed algorithm achieved high accuracy (>90%) on two independent EEG datasets.
• Optimized DWT settings were found to substantially impact seizure detection performance.
• The method demonstrated superior accuracy and transferability across different datasets compared to existing wavelet-based techniques.

Conclusions:

• Automated optimization of DWT settings provides a more accurate and computationally efficient approach to epileptic seizure detection from EEG.
• The developed framework offers a robust and transferable solution for computer-aided seizure detection.
• This study highlights the critical role of parameter optimization in DWT-based EEG signal analysis.