Aperiodic-Aware EEG Features for Cross-Subject Acute Sleep Deprivation Detection

Abstract

. Acute sleep deprivation impairs vigilance and cognitive stability. Resting-state electroencephalography (EEG) is a potential approach, although traditional bandpower features may be confounded by broadband aperiodic activity. In this study, the authors assessed whether aperiodic-aware EEG characteristics could serve as useful indicators for detecting acute sleep deprivation across subjects. Eyes-open resting-state EEG recordings from the OpenNeuro ds004902 dataset were analyzed. Within-subject statistical tests demonstrated that the aperiodic exponent increased under sleep deprivation in both reduced-channel and all-channels configurations and that this trend persisted in a strict sensitivity subset. Neither session order nor age was significantly associated with the primary exponent change. By contrast, cross-subject discrimination with lightweight classifiers was only moderate, implying that generalization remains difficult despite the repeatability of the physiological change. These results provide a feasible basis for future EEG-based monitoring analyses.