Abstract—An EMD-chaos based approach is proposed to discriminate EEG signals corresponding to healthy persons, and epileptic patients during seizure-free intervals and seizure attacks. An electroencephalogram (EEG) is first empirically decomposed to intrinsic mode functions (IMFs). The nonlinear dynamics of these IMFs are quantified in terms of the largest Lyapunov exponent (LLE) and correlation dimension (CD). This chaotic analysis in EMD domain is applied to a large group of EEG signals corresponding to healthy persons as well as epileptic patients (both with and without seizure attacks). It is shown that the values of the obtained LLE and CD exhibit features by which EEG for seizure attacks can be clearly distinguished from other EEG signals in the EMD domain. Thus, the proposed approach may aid researchers in developing effective techniques to predict seizure activities.

Index Terms—Electroencephalogram (EEG), empirical mode decomposition (EMD), largest Lyapunov exponent (LLE), correlation dimension (CD), epileptic seizures.

The Authors are with the Electrical and Electronic Engineering Department, Bangladesh University of Engineering and Technology,Dhaka-1000, Bangladesh (e-mail: imamul@eee.buet.ac.bd)

Cite: S. M. Shafiul Alam,S. M. Shafiul Alam, Aurangozeb, and Syed Tarek Shahriar, "EEG Signal Discrimination using Non-linear Dynamics in the EMD Domain," International Journal of Computer and Electrical Engineering vol. 4, no. 3, pp. 326-330, 2012.