Comparison of a Set of Power Distribution Parameters for Intraoperative EEG Monitoring

This study was conducted to compare the relative performance of 19 univariate EEG power distribution parameters in terms of their sensitivity to changes in the background (stationary) EEG activity and insensitivity to broad-band interference. Each of the parameters partitioned the power spectrum into two frequency subbands, each subband containing a fixed specific percentage of the total power. The EEG signals recorded from ten patients during induction with thiopental sodium were used in the study. Two sensitivity functions were defined to compare the sensitivity of the parameters to the effects of thiopental sodium. Parameter changes were compared to bandwidth changes during induction to evaluate their sensitivity to broad-band interference. The results of the study showed that the parameters in the middle of the percentile scale were more sensitive to background EEG changes produced by thiopental sodium than were those at either end of the scale. Sensitivity to broad-band interference was found to be directly proportional to percentile. Within the limits of this study, a parameter between the 25th and 30th percentiles would be the most suitable parameter to monitor intraoperatively.