Proper diagnosis of epilepsy requires the detection and analysis of epileptic seizures. Manual monitoring of long term EEG is tedious and costly. Therefore, a reliable automated seizure detection system is desirable. Most current state-of-the-art methods use hand crafted feature extraction and simple classification techniques. In this study, we have achieved better accuracy introducing two novel approaches. A Deep recurrent neural network (DRNN) architecture that performs automated patient specific seizure detection using scalp EEG. A unique mapping of seizure EEG signal for efficient processing with the DRNN allows the proposed deep architecture to simultaneously learn both temporal and spatial features of raw seizure EEG respectively. Overall, the proposed network successfully detects 100% of total seizure events with an average with low detection delay. The results demonstrate superior performance to that of the current state-of-art seizure detection methods. The proposed DRNN architecture also has low processing time with sparse use of computing resources and superior performance make the proposed architecture appropriate for real-time use.
Real Time Network Response of Seizure Detection Network (seizure event from 4920 sec to 5006 sec)