A Fusion of CNN and Grey Wolf Optimization-Enhanced BiGRU for Epileptic Seizure Recognition Using EEG Signals

Epilepsy, a profoundly impacting neurological disorder, afflicts an estimated 65 million individuals across the globe. Early detection of seizures, a critical aspect of epilepsy management, poses an ongoing challenge. This study introduces an innovative hybrid deep learning approach that integrates Convolutional Neural Networks (CNN) for automated feature extraction with Grey Wolf Optimization-based Bidirectional Gated Recurrent Unit (BiGRU) networks to differentiate between three fundamental stages of epileptic seizures: preictal, interictal, and ictal, utilizing Electroencephalogram (EEG) signals. The Grey Wolf Optimization algorithm is employed to determine the most suitable hyperparameters for configuring the BiGRU network. The research seeks to address this pressing need, ultimately promising faster and more effective treatment, which is essential for improving the lives of those with epilepsy. Importantly, the proposed method significantly advances the field of epileptic seizure detection by achieving an accuracy of 98.67%. The obtained Precision, F1-score, and Recall values are 98%, 97%, and 98% respectively.