Advanced Epileptic Seizure Recognition with a Hybrid CNN-BiLSTM Model on EEG Signals
L. Krishna Kumari, R Prem Ananth, K Ramalakshmi, Balaji Mohan, S. P. Santhoshkumar, S. Mahalakshmi
Abstract
Epileptic seizure recognition (ESR) is a vital area in healthcare, neurology, and patient monitoring, particularly for managing epilepsy in real time. While deep learning (DL) models have shown exceptional performance in recognizing seizures from EEG signals, processing sequential and time-series data remains challenging. This study proposes a hybrid deep learning framework combining one-dimensional convolutional neural networks (1D CNN) and bidirectional long short-term memory (BiLSTM) networks to enhance seizure recognition. The 1D CNN efficiently captures localized temporal features from EEG data, while the BiLSTM models bidirectional long-range dependencies, improving temporal feature representation. Tested on a benchmark dataset, the model achieves an outstanding accuracy of 98.18%, surpassing existing state-of-the-art methods. These results underscore the effectiveness of integrating CNN for feature extraction and BiLSTM for sequence modeling, offering a robust and reliable approach for automated ESR in clinical applications.