EEG Sensor-Based Frequency Domain Analysis for Epileptic Seizure Detection

Abstract

FMRI cannot apply to patients with implants and it has several limitations like temporal resolution, hence the EEG based detection of abnormal activities for epileptic patients is very important as if correct treatment and medication given to patient on correct time it can make the patient seizure free. In this paper we have proposed technique and it’s outcomes for the detection of epileptic seizure by using frequency domain analysis technique of EEG sensor array. And comparing it we’ve achieved classification accuracy is less 59% which is as compared to our previous study using time domain features which was 96.5%. We conclude that frequency domain features like Power spectral density, Kurtosis shall not be the selection but time domain features such as standard deviation and quartile ranges shall be considered.