MEASURING THE DISSIMILARITY BETWEEN EEG RECORDINGS THROUGH A NONLINEAR DYNAMICAL SYSTEM APPROACH

A new measure of dissimilarity between two EEG segments is proposed. I t is derived from the application of the mathematical concept of dista nce between series of one-step predictions according to the estimated non-linear autoregressive functions. The non-linear autoregressive est imation is performed by non-parametric regression using kernel estimat ors. The possibility of applying this measure for automatic classifica tion of EEG segments is explored. For this purpose multidimensional sc aling anti cluster analyses are applied on the basis of the calculated dissimilarity measures. In particular, its application to different E EG segments with delta activity and also with alpha waves reveals high agreement with visual classification by EEG specialists.