Relation of frequency-analyzed EEG to monitoring behavior

Abstract

Two experiments examined the relation of prestimulus electroencephalographic (EEG) frequency to choice RT. Based on previous studies, initial analysis was concerned with 3--7 c/sec activity. In well-rested subjects (Exp. 1), trial-by-trial analyses indicated large variations in prestimulus EEG activity which were unrelated to RT and large variations in RT which were unrelated to prestimulus EEG. In Exp. 2, subjects were deprived of sleep for 1 night and within-subject comparisons made between RT and EEG activity immediately preceding the 10 shortest and 10 longest RT trials, and 10 trials where the subject failed to respond. Significant univariate correlations were found, largely between RT and the frequencies in the 15--20 c/sec range of EEG activity and not in the 3--7 c/sec activity. A multiple regression analysis using up to 5 EEG frequencies indicated significant correlations of prestimulus EEG activity with RT; but with considerable subject-to-subject variability in the EEG frequencies contributing to the multiple R. The overall results suggest that there can be considerable variation in EEG activity which is unrelated to performance, except when the EEG fluctuations are secondary to changes in arousal which, in turn, effect performance.