A Simple Wavelet Method for Automated Detection of Epileptic Neural Spikes in Electroencephalogram

Abstract

In clinical diagnosis process of epilepsy, the physicians rely on tedious visual screening of long time recorded electroencephalograms (EEG) for detection and localization of important epileptic events. This process is very time consuming and there is also possibility of making wrong diagnosis owing to tiredness or to subjectivity of the specialist. Therefore, the computer-assisted diagnosis systems play an important role in quick and correct diagnosis of epilepsy for further medical treatment of the patients. In this paper we propose a new method of a decision-making tool supported by experimental data to detect ictal events in EEG signals. Our method is based on continuous wavelet transform (CWT) with suitable mother wavelet functions and thresholding technique. We demonstrate the efficiency of our method on data to identify and clearly locate in time the seizure activities. The method is superior both in denoising and in identifying any abnormal epileptic spikes based on in sets of combined scales. Moreover, our method is fast and simple which is also computationally efficient