On-line Methodology for Detection of Suspected Epileptic Seizures from Electroencephalogram (EEG)

Abstract

An on line methodology is introduced for detecting suspected seizures from the electroencephalogram(EEG). This is achieved by detecting the nonstationarities of EEG. Two seizure indices, namely; γ(t) and α(t) where [0≤α(t),γ(t)≤1], are introduced. In the absence of seizure, EEG remains stationary and hence seizure indices α(t) and γ(t) remain unity. If and when abnormality or epileptic seizure activity occurs, the EEG becomes nonstationary causing the seizure indices to drop in value. The dips in the seizure indices indicate the time at which seizure activity occurs while the strength of the dips indicates the severity of the seizure activity. If the seizure indices drop below a preset threshold value this could activate a pacemaker as a preventive measure of epileptic seizure. Methodology is not limited to epileptic seizure detection and can be used to detect malfunctions of stationary systems in general.