Saturation of electroencephalogram permutation entropy for large time lags

New approach to investigation of electroencephalogram (EEG) patterns over different time spans is proposed based on permutation entropy (PE) measure. PE for large time lags (1-100 samples) and orders from 2 to 8 was calculated for EEG signals of three types: EEG from healthy subject, EEG containing high-magnitude bursts and EEG with epileptiform complexes. PE value saturation phenomena is revealed for large time lags (approximately 10 or 30 samples for different EEG types) and for whole range of orders used in this study. This effect might be due to the inherent EEG property, namely that the samples in each EEG pattern are spaced too far from each other and they are like samples from some arbitrary stochastic process.