Deep Learning Based Approach for Epilepsy Detection Using EEG Data

Abstract

Epilepsy is one of the most common neurological disorders characterized by recurrent seizures. Electroencephalography (EEG) serves as a crucial diagnostic tool for epilepsy, yet traditional diagnosis relies heavily on manual interpretation, which is time-intensive, subjective, and prone to errors. This study addresses the need for automated, reliable, and efficient diagnostic methods by exploring the classification of healthy and epileptic individuals using raw EEG data analyzed through a one-dimensional Convolutional Neural Network (1D CNN). The proposed model was trained and evaluated on a dataset comprising 148 scalp EEG recordings (72 epileptic patients and 78 healthy individuals) obtained from a local hospital. The CNN model automatically extracted features from the EEG signals and achieved an accuracy of 97.73%, sensitivity of 98%, and specificity of 98%. Channel-specific analyses were conducted to evaluate the contribution of individual EEG channels, and the model's performance was further examined by progressively reducing the number of channels. These findings underscore the potential of integrating EEG data with deep learning for accurate, automated, and non-invasive epilepsy diagnosis. Additionally, the study highlights the significance of channel reduction in simplifying data while preserving diagnostic precision, facilitating more efficient clinical applications and real-time seizure detection systems.