Beamforming aided fast fourier transform to detect epileptic seizure onset based on scalp EEG signals and different learning methods

Abstract

Epilepsy is a neurological disorder that causes serious seizure attacks. Effect of epileptic seizure is now so severe in world that it causes patients lifetime phenomenon which can be hindrance to people to get back to normal life again. Electroencephalogram (EEG) signals help to understand brain's activity more insight-fully. These EEG data recorded from normal and seizure patient to understand the brain activity. For classifying ictal and interictal period most of the classifier do not give 100 percent accuracy. In this paper, EEG signals has been used to analysis the distribution time-frequency features than calculated energy feature from it using Fourier transform methods. Beamforming methods using as noise distortion to have less noise to get more accuracy. The features are used for training data is in machine learning. Support vector machine is used as a classifier and this approach give us 89.8 percent accuracy for detecting epileptic seizures.