Automatic detection of epileptiform discharges in EEG using a back-propagation network

This paper presents an automatic approach to detect epileptiform discharges (ED) in electroencephalogram (EEG). On the algorithm we utilized back-propagation artificial neural network (BPN) to detect ED. We train BPN respectively for each patient and induce parameter k to determine a threshold value. The result shows that the algorithm can determine presence or absence of ED automatically, and decrease the false determination in current automated approaches as well.