{"ID":2850424,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.02846","arxiv_id":"2511.02846","title":"Spatio-Temporal Attention Network for Epileptic Seizure Prediction","abstract":"In this study, we present a deep learning framework that learns complex spatio-temporal correlation structures of EEG signals through a Spatio-Temporal Attention Network (STAN) for accurate predictions of onset of seizures for Epilepsy patients. Unlike existing methods, which rely on feature engineering and/or assume fixed preictal durations, our approach simultaneously models spatio-temporal correlations through STAN and employs an adversarial discriminator to distinguish preictal from interictal attention patterns, enabling patient-specific learning. Evaluation on CHB-MIT and MSSM datasets demonstrates 96.6\\% sensitivity with 0.011/h false detection rate on CHB-MIT, and 94.2% sensitivity with 0.063/h FDR on MSSM, significantly outperforming state-of-the-art methods. The framework reliably detects preictal states at least 15 minutes before an onset, with patient-specific windows extending to 45 minutes, providing sufficient intervention time for clinical applications.","short_abstract":"In this study, we present a deep learning framework that learns complex spatio-temporal correlation structures of EEG signals through a Spatio-Temporal Attention Network (STAN) for accurate predictions of onset of seizures for Epilepsy patients. Unlike existing methods, which rely on feature engineering and/or assume f...","url_abs":"https://arxiv.org/abs/2511.02846","url_pdf":"https://arxiv.org/pdf/2511.02846v1","authors":"[\"Zan Li\",\"Kyongmin Yeo\",\"Wesley Gifford\",\"Lara Marcuse\",\"Madeline Fields\",\"Bülent Yener\"]","published":"2025-10-24T01:45:25Z","proceeding":"eess.SP","tasks":"[\"eess.SP\",\"cs.AI\",\"cs.LG\"]","methods":"[]","has_code":false}