Analysis of EEG Signals Using Empirical Mode Decomposition and Relative Band Energy

This paper approaches an analysis of various epileptiform EEG signals using Empirical Mode Decomposition in the time domain and relative band energy distribution, in the frequency domain. Applying EMD, various EEG waveforms are decomposed into a set of 8 intrinsic mode functions, then to each of them FFT is applied, and finally their relative band energy distribution is obtained. A statistical analysis of the instantaneous frequency values of each IMF is also performed. Thus, we study the possibility to describe a given EEG pattern by relative band energy distribution of its components given by EMD analysis.