EEG independent component and coherence analysis from scalp and depth electrodes during propofol anesthesia

Abstract

Independent Component Analysis (ICA) and Coherence Analysis are used to separate and follow the behavior of the components of EEG burst suppression pattern in propofol anesthesia. The signal was measured from 13 scalp electrodes and 4 depth electrodes implanted for the treatment of parkinsonism. The results show that ICA is capable in separating the components (spindle, mixed frequency activity and sharp wave) implying that the components can be statistically described. Coherence analysis shows that the beta activity, corresponding to spindles, starts actually several minutes before the onset of burst suppression and slows down gradually. The results indicate that there are patterns in EEG during propofol anesthesia, which have not been described previously. These patterns should give insight into the mechanisms of unconsciousness during anesthesia and provide new approaches to using EEG in pharmacological studies.