We explored brain functions accompanying cognitive load and affective arousal by examining task accuracy, reaction time, time series EEG power of the alpha band, and estimation of arousal level using a machine learning model based on EEG data, during nback tasks. Increased variability of reaction time and fluctuation of the alpha power at 0.01 Hz were shown in a 3-back task compared to a 0-back task. Enhancement of the alpha power fluctuation at the very low frequency related to higher task performance, and less arousal level estimated by the model, suggesting ability of the brain to flexibly adjust behaviors and arousal level to cognitive load.
