EEG-Based Patient Independent Epileptic Seizure Detection Using GCN-BRF

Epilepsy affects an estimated 50–70 million people worldwide, making it one of the most prevalent neurological disorders. Detecting seizures, a primary symptom of epilepsy, poses significant challenges due to the limitations of patient-specific models and the complex nature of EEG signal analysis. Traditional methods, which focus on isolated EEG channels, often fail to capture the dynamic interconnections within the brain's network that are essential for accurate seizure detection. This study introduces a novel, patient-independent approach using Graph Convolutional Networks (GCNs) to process graph-structured data representing the brain's network. Our methodology employs a comprehensive feature set derived from Random Forest selection, encompassing 37 node-specific features and two global features: the GCN's classification output and an eigenvector from the correlation matrix. This rich feature representation allows for an in-depth analysis of the structural properties of EEG data. The proposed model was evaluated on unseen data from four patients and demonstrated exceptional generalizability and performance, achieving notable metrics such as 91.70% accuracy, 91.32% precision, 88.71% sensitivity, and a 91.57% F1-Score in patient-independent settings. Further patient-specific evaluations reinforced the model's efficacy, with near-perfect scores across all key metrics. Our findings highlight the potential of GCNs to overcome existing challenges in seizure detection, offering a promising direction for epilepsy care and contributing valuable insights into the analysis of neurological disorders.