Interictal epileptic discharge EEG detection based on wavelet and multiresolution analysis

Abstract

Epileptologists use interictal epileptic discharge (lED) as a marker for epilepsy. The present conventional method to distinguish normal and I ED by an epileptologist's visual screening is tedious and operator dependent. The focus of this paper is to distinguish normal and IED in clinically recorded electroencephalogram (EEG) using discrete wavelet transform. Wavelet multiresolution analysis has been adopted in this study looking into wavelet energy, wavelet entropy and amplitude dispersion in every sub-band. The extracted features were classified using support vector machine (SVM). EEG data were obtained from both online database and Pusat Perubatan Universiti Kebangsaan Malaysia (PPUKM) Neurology database. The ability of the proposed algorithm in detecting the presence of IED is 96.5% of accuracy, 100% of sensitivity and 95.5% of specificity. The algorithm has good potential to be used in clinical practice for IED detection with validation against the present clinical detection method.