A neural signature of sleep deprivation in the human brain

Abstract Insufficient sleep disrupts cognitive and emotional functioning, yet the precise neural consequences of sleep loss and their persistence remain unclear. Here, we leverage machine learning and large neuroimaging datasets to identify a candidate neural signature that robustly distinguishes sleep-deprived from well-rested brains. We validate this signature across multiple independent datasets spanning both controlled experimental and real-world settings. The signature not only detects residual neural disturbances following a night of recovery sleep, but also demonstrates sensitivity to partial sleep deprivation. Additionally, it captures natural variations in sleep duration in the general population, independent of experimental manipulation. We further identify distributed connectivity patterns that contribute to the signature, highlighting networks vulnerable to sleep manipulations and those that are resistant or rapidly normalized after recovery sleep. The reliability and generalizability of this neural signature underscore its potential as a biomarker for understanding and monitoring the neural impacts of acute and chronic sleep loss.