Explainable automated seizure detection using attentive deep multi-view networks

Aref Einizade, Samaneh Nasiri, Mohsen Mozafari, Sepideh Hajipour Sardouie, Gari D. Clifford

Abstract: Highlights•Manual inspection of epileptic seizures is time-consuming and prone to inter-rater variability.•We propose an artifact rejection approach to remove unwanted signals.•We also propose a fusion attentive deep multi-view network (fAttNet) for seizure detection.•Our proposed approach is successfully applied to a cross-subject scenario.•More importantly, the proposed method is interpretable for medical professionals.