Six-center Assessment of CNN-Transformer with Belief Matching Loss for Patient-independent Seizure Detection in EEG
Abstract
Neurologists typically identify epileptic seizures from electroencephalograms (EEGs) by visual inspec-tion. This process is often time-consuming, especially for EEG recordings that last hours or days.To expedite the process, a reliable, automated, and patient-independent seizure detector is essential.However, developing a patient-independent seizure detector is challenging as seizures exhibit diversecharacteristics across patients and recording devices. In this study, we propose a patient-independentseizure detector to automatically detect seizures in both scalp EEG and intracranial EEG (iEEG). First,we deploy a convolutional neural network with transformers and belief matching loss to detect seizuresin single-channel EEG segments. Next, we extract regional features from the channel-level outputs todetect seizures in multi-channel EEG segments. At last, we apply postprocessing filters to the segment-level outputs to determine seizures’ start and end points in multi-channel EEGs. Finally, we introducethe minimum overlap evaluation scoring as an evaluation metric that accounts for minimum overlapbetween the detection and seizure, improving upon existing assessment metrics. We trained the seizuredetector on the Temple University Hospital Seizure (TUH-SZ) dataset and evaluated it on five indepen-dent EEG datasets. We evaluate the systems with the following metrics — sensitivity (SEN), precision(PRE), and average and median false positive rate per hour (aFPR/h and mFPR/h). Across four adultscalp EEG and iEEG datasets, we obtained SEN of 0.617-1.00, PRE of 0.534-1.00, aFPR/h of 0.425-2.002, and mFPR/h of 0-1.003. The proposed seizure detector can detect seizures in adult EEGs andtakes less than 15s for a 30 minutes EEG. Hence, this system could aid clinicians in reliably identifyingseizures expeditiously, allocating more time for devising proper treatment.
