0213 Network-Level Brain Connectivity Predicts Resilience to Sleep Deprivation

Abstract Introduction Sleep deprivation (SD) reliably impairs vigilance, but people differ considerably in how quickly and severely their performance declines. Understanding the neural patterns that accompany this variability may help identify mechanisms of resilience to SD. Here we examined whether the Resilience to Sleep Deprivation Metric (RSDM), a behavioral summary of preserved reaction speed relative to baseline, was reflected in the way functional brain networks reorganize during prolonged wakefulness. Methods Sixteen healthy adults (8 females, age = 23.6 ± 4.6) completed a 39-hour SD protocol with six resting-state fMRI scans collected every six hours. Psychomotor vigilance task (PVT) performance was assessed at each scan, and RSDM was computed as the ratio of reaction speed during later (Sessions 4-6) versus earlier (Sessions 1-3) phases. Functional connectivity was estimated across 375 regions (Glasser cortical + Harvard-Oxford subcortical). Network-Based Statistics (NBS) tested RSDM × Hours Awake interactions on edge-level connectivity (FWE-corrected p < .05), controlling for baseline lapses, circadian phase, and random effects. Nodal strength was evaluated for regions identified in NBS, with FDR correction applied across these nodes. Results Two subnetworks showed significant resilience-dependent modulation (FWE-corrected p < .05). One involved bilateral thalamus, globus pallidus, and posterior cingulate -- regions linked to arousal and thalamocortical relay. The other included ventral and dorsal visual cortices, auditory areas, and medial temporal regions involved in perception and memory. In both networks, individuals with higher RSDM values (i.e., resilient) maintained or increased connectivity as waking time accumulated, whereas those with lower RSDM (i.e., Vulnerable) showed progressive reductions. Nodal strength analyses followed the same pattern, highlighting resilience-related differences in thalamus, basal ganglia, and ventral visual cortex (FDR-corrected p < .05). Conclusion Resilient individuals appear to preserve integration within thalamocortical and perceptual-memory networks during extended wakefulness. These network features may help clarify why some individuals are able to sustain performance during sleep loss and could inform future efforts to develop imaging-based markers or strategies to mitigate fatigue-related decline. Support (if any) Supported by Army Research Office award W911NF2210223, subaward SUBK00016417.