Computerized Analysis of EEG to Determine Focal Epilepsy

Abstract

Epilepsy is a neurological disorder that occurs due to the abnormal electrical discharges in the brain, thus affecting the patient's personality, behavior and day-to-day routine. Epilepsy can be broadly classified into generalized epilepsy and focal epilepsy. Focal epilepsy occurs due to abnormal electrical discharges in the smaller section of the brain. These electrical discharges later spread to the larger part of the brain, resulting in generalized epilepsy. In this paper, a computer-aided diagnostic system is proposed to detect the focal epilepsy using electroencephalogram (EEG) signals. Features from the decomposed signals are extracted using fuzzy approximation entropy, Higuchi's fractal dimension and correlation dimension. Classification accuracy of 87.68% and 84.45% was obtained using K nearest neighbor (KNN) and support vector machine (SVM) classifiers respectively.