Normal inverse Gaussian parameters in the empirical mode decomposition domain for the detection of epilepsy and seizure

Abstract

In this paper, a comprehensive analysis of electroencephalogram (EEG) signals is carried out in the empirical mode decomposition (EMD) domain using a publicly available benchmark EEG database. First, the intrinsic mode functions (IMF) are extracted in the EMD domain. Next, normal inverse Gaussian (NIG) probability density function (pdf) is introduced and it is investigated whether the NIG pdf can suitably model the IMFs extracted in EMD domain of the EEG signals. It is shown that the NIG pdf is a suitable prior to model the first five IMFs extracted from various types of EEG recordings. It is further shown that the NIG parameters can distinguish among the EEG signals at the five IMF levels quite well. The analysis is further confirmed through the p-values obtained by one way ANOVA analysis. Thus, the NIG parameters in the EMD domain may be used to characterize EEG signals and help the researchers in developing fast, effective and improved classifiers for the detection of epilepsy and seizure.