Higher-Order Statistics of Stockwell Transform for Epileptic Seizure Detection from EEG Signals

Epilepsy is one of the most common neurological disorders and its characteristic is recurrent sudden abnormalreactions of brain happens when the neurons generate abnormal electrical discharges from brain cells.Electroencephalography (EEG) signal is used for diagnosis of electrical activity of brain. In this paper, we present anefficient algorithm for epileptic detection based on time-freqency analysis of EEG signals. After computing Stockwelltransfrom of EEG signal, higher-order statistics such as cumulants are computed from its sub-bands. Multi-clusterfeature selection (MCFS) is used to select informative features and after that supprt vector machine (SVM) is used forclassification. Results demonstrate the efficiency of the proposed method in epileptic seizure detection.

Keywords:

Cumulant , Epileptic seizure , feature selection , Stockwell transform