Abstract
INTRODUCTION: Disorders of consciousness (DoC), including unresponsive wakefulness syndrome (UWS) and minimally conscious state (MCS), are primarily diagnosed behaviorally. Recent evidence indicates that loss of consciousness manifests as irregularities in neural oscillatory activity across delta, theta, and alpha frequency bands. However, conventional spectral analysis often conflates periodic oscillations with aperiodic 1/f components, potentially obscuring consciousness-related dynamics. METHODS: To elucidate the mechanistic basis of spectral alterations in consciousness impairment, we compared oscillatory and aperiodic activity patterns in the electroencephalogram (EEG) of patients with different consciousness levels. We further examined the spatiotemporal variability of these neural signatures and rigorously evaluated their discriminative power for state classification using support vector machine (SVM) analysis. RESULTS: While periodic and aperiodic activities are independent, our results indicate that both components exhibit significant differences between groups at both local and global scales. Critically, higher spatial and temporal variability of aperiodic features (spectral exponent) were correlated with preserved consciousness. When distinguishing UWS from MCS, the combination of periodic and aperiodic features significantly improved classification performance compared to using either metric alone. DISCUSSION: Our findings demonstrate that both periodic oscillations and aperiodic activity provide valuable information about consciousness levels. Critically, the spatiotemporal dynamics of the aperiodic component serve as a key marker of brain state. This underscores the necessity of accounting for aperiodic activity in mechanistic studies and clinical assessments of DoC.