Nonlinear Dynamic Analysis Of EEG In Emotional Events

The emotional evaluation system is an automatic recognition of emotional systems; it’s an important issue in the human-machine interaction engineering. EEG-based emotion evaluation system for its reliable, convenient features, showed a rising trend in recent years.This paper attempts to use nonlinear dynamics methods to extract characteristics of pleasant and unpleasant emotions from EEG, and to find optimal characteristics to distinguish these two basic emotions, through the analysis of the nonlinear dynamics of the EEG. research the correlation of emotion and EEG.EEG data was recorded either from16electrodes with a standard EEG-cap according to the10-20system. Subjects are20male college students (mean age22years), emotions induced by music audios (eyes closed and relax) Calculation of the parameters of the LZC(Lempcl-Ziv Complexity)and ApEn(Approximate Entropy) analysis of brain waves, to extract the nonlinear characteristics of the brain waves of different emotions.By calculating the EEG of two basic emotions, showed LZC and ApEn robust correlated with emotion. Alpha rhythm’s LZC is positively correlated with valence, while the alpha rhythm’s ApEn is negatively correlated with valence. The statistical analysis result showed that EEG rhythm and emotional state can significantly affect the LZC and ApEn (p<0.05). LZC of the significant differences (p<0.05) reflected in the Alpha rhythm Fp1, FP2, P3, O2, F7, F8, T3, T6electrode, the Beta rhythm of the T5electrode, the Theta rhythm of the F3, O1electrode; the ApEn of the significantly with the difference Theta rhythm Fpl and P3electrodes.LZC and ApEn both are robust correlation with emotion, they can better distinguish pleasant and unpleasant emotion, but we need to notion the influence of rhythm and electrode when using these two methods to extract EEG characters.