Analysis of electroencephalogram for the recognition of epileptogenic area using ensemble empirical mode decomposition

Abstract

Recognizing the epileptogenic area of a brain is done by analyzing the electroencephalogram signal. This area is responsible for the occurrence of seizure activity in a brain. In this paper, a methodology has been presented for the analysis of electroencephalogram to recognize epileptogenic area of brain. Ensemble empirical mode decomposition (EEMD) has been used for the estimation of intrinsic mode functions (IMFs), and six parameters consisting of statistical and frequency-based feature have been extracted from first ten IMFs. The ReliefF algorithm has been used to select the relevant features for the training of artificial neural network (ANN) for recognition of epileptogenic area. The methodology has been evaluated based on accuracy, specificity and sensitivity. The comparison has also been made with other methods of epileptogenic area detection where it has been observed that the proposed method outshines other.