Method, apparatus and computer program product for automatic seizure monitoring

Abstract

Method, apparatus and computer program product for monitoring seizure activity in brain are disclosed. At least one parameter set time series is derived from brain wave signal data obtained from a subject, wherein each parameter set sequence comprises sequential parameter sets and each parameter set comprises values for at least two signal parameters, the values being derived from the brain wave signal data. In order to reduce susceptibility to inter-subject variations and to enhance adaptability to each recording, past EEG signal data of the subject is used to determine an envelope object that encompasses the parameter points that sequential parameter sets derived from the past signal data form in a parameter space. A reference point is also determined, whose location in the parameter space depends on the past signal data. At least one new parameter point is then obtained from the subject and an evolution indicator set is determined. By examining whether the evolution indicator set fulfills predetermined location and direction criteria in relation to the envelope object and the reference point, seizure activity may be detected. The envelope object and the reference point are conditionally updated for on-line measurement.