The Neural Correlates of Consciousness: A Spectral Exponent Approach to Diagnosing Disorders of Consciousness.

Abstract

Disorder of consciousness (DoC) poses diagnostic challenges due to behavioral assessment limitations. This study evaluates the spectral exponent (SE)-a neurophysiological biomarker quantifying the decay slope of electroencephalography (EEG) aperiodic activity-as an objective tool for consciousness stratification and clinical behavior scores correlation. The study involved 15 DoC patients, nine conscious brain-injured controls (BI), and 23 healthy controls (HC). Resting-state 32-channel EEG data were analyzed to compute SE across broadband (1-40 Hz) and narrowband (1-20 Hz, 20-40 Hz). Statistical frameworks included Bonferroni-corrected Kruskal-Wallis H tests, Bayesian ANOVA, and correlation analyses with CRS-R behavioral scores. Narrowband SE (1-20 Hz) showed superior diagnostic sensitivity, differentiating DoC from controls (HC vs. DoC: p < 0.0001; BI vs. DoC: p = 0.0006) and MCS from VS/UWS (p = 0.0014). SE correlated positively with CRS-R index (1-20 Hz: r = 0.590, p = 0.021) and visual subscale (1-20 Hz: r = 0.684, p = 0.005). High-frequency (20-40 Hz) SE exhibited inconsistent results. Longitudinal tracking in an individual revealed a reduction in SE negativity, a flattening of the 1/f slope, and behavioral recovery occurring in parallel. Narrowband SE (1-20 Hz) is a robust biomarker for consciousness quantification, overcoming behavioral assessment subjectivity. Its correlation with visual function highlights potential clinical utility. Future studies should validate SE in larger cohorts and integrate multimodal neuroimaging.