NONLINEAR FORECASTING MEASUREMENTS OF THE HUMAN EEG DURING EVOKED EMOTIONS

Forecasting ability applied to a single EEG time sequence may quantify loss of memory of past brain states. This fundamental property of com plex dynamical systems could prove to be an important measure of curre nt brain states. The present study examined nonlinear forecasting (or nonlinear predictability) estimates of the scalp-recorded EEG in 76 he althy volunteers in response to emotionally charged (negative, positiv e and control neutral) video-clip stimuli. EEG was recorded from 18 si tes (10-20 system). The obtained results indicate that the negative em otional impact led to a more predictable EEG dynamics, compared with n eutral or positive emotional video content, and this effect was restri cted to the posterior cerebral sites. The studied signatures of EEG dy namics succeed also in specific discrimination between effects of posi tive and negative film categories: significantly more predictable dyna mics over posterior cerebral sites in response to negative film catego ry was accompanied by enhanced predictability in answer to positive fi lm category over frontocortical loci. It is concluded that these findi ngs suggest the association of basic cortical nonlinear mechanisms wit h the specific physiological processes of emotional processing.