Epileptic seizure detection from electroencephalogram signals based on 1D CNN-LSTM deep learning model using discrete...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Artificial Intelligence

Background:

• Epileptic seizures result from excessive brain electrical activity, detectable via Electroencephalogram (EEG) signals.
• Automated seizure detection from EEG is crucial for diagnosis and patient monitoring.
• Existing methods may not fully capture the complex spatiotemporal dynamics of EEG signals.

Purpose of the Study:

• To develop and evaluate an automated system for epileptic seizure identification using EEG signals.
• To leverage deep learning models, specifically CNN and LSTM, for enhanced feature extraction from EEG data.
• To compare the proposed model's performance against existing machine learning classifiers.

Main Methods:

• EEG signals were processed by extracting and concatenating frequency bands using Discrete Wavelet Transform (DWT).
• A 1D Convolutional Neural Network (CNN) was employed to extract spatial features from the EEG data.
• A Long Short-Term Memory (LSTM) layer processed the CNN output to capture temporal dependencies, followed by a fully connected layer for classification.

Main Results:

• The proposed model achieved high accuracy on multiple datasets: TUSZ corpus (94.32%), BONN (97.24%), and CHB-MIT (96.94%).
• Performance metrics including Kappa values and GDR demonstrated the model's effectiveness across different EEG datasets.
• The model significantly outperformed several popular machine learning classifiers in seizure detection accuracy.

Conclusions:

• The integrated CNN-LSTM model effectively extracts spatiotemporal features from EEG signals for accurate seizure detection.
• The proposed method offers a robust and high-performing solution for automated epileptic seizure identification.
• The CNN's capability in extracting spatial features is a key factor contributing to the model's superior performance.