Real-time detection of epileptogenic discharges using autoregressive prediction

An algorithm is presented that detects epileptogenic discharges from the electroencephalograph (EEG) and magnetoencephalograph (MEG) signals of an epileptic patient in real time using autoregressive prediction in conjunction with an adaptive binomial decision rule. This algorithm is implemented on a personal computer system that performs the data acquisition and discharge detection and extraction functions. Testing with both simulated and real patient data shows this method to be useful for real-time epileptogenic discharge detection. >