Reduced but Reversible Brain Entropy After Occupational Partial Sleep Deprivation in Night-Shift Medical Staff

Partial sleep deprivation (PSD) poses health risks to the night-shift workers (NSW), but its underlying impacts on local brain function remain underexplored. Brain entropy (BEN), a nonlinear dynamic metric, has emerged as a novel parameter of choice in probing the temporal irregularity of brain activity, thus may offer new insights into the characterization of the brain dysfunction following sleep deprivation.

Seventy-eight female medical NSWs and 30 non-NSW healthy controls (HC) were recruited in this study. Psychomotor vigilance tasks (PVT) and resting-state fMRI (rs-fMRI) were sequentially conducted at three conditions for NSWs: baseline (prior to NSW), PSD (immediately after NSW), and recovery (after 3-5 days of regular sleep). Nine NSWs were excluded due to consecutive or intermittent sleep duration exceeding 6 h on the night-shift day, or significant head motion in the rs-fMRI scans, resulting in 69 NSWs in the following analysis. Static and dynamic sample entropy (SampEn) metrics were calculated for BEN quantification. An L1-regularized logistic regression (LR) model was constructed based on baseline SampEn metrics to distinguish PSD-vulnerable (N = 35) and PSD-resistant (N = 34) individuals with their PVT performances as the classification reference.

Compared to HCs, SampEn of young female NSWs reduced mainly in the occipital and temporal cortices, and associated with poor sleep quality. But these inter-group differences did not survive the strict multiple comparison correction. Both static and dynamic SampEn metrics of NSWs were significantly altered following acute PSD, and largely returned to the baseline after sleep recovery. The SampEn-based LR model achieved a classification accuracy value of 78.26% to distinguish PSD-vulnerable and PSD-resistant individuals.

One-night of acute PSD in shift work leads to reduced but largely reversible BEN, whereas long-term occupational chronic PSD tends to reduce the BEN in distributed primary cortices and correlates with the poor sleep quality in young female NSWs. The BEN metrics could serve as a potential biomarker to identify individual susceptibility to PSD, which may contribute to the optimization of rotating-shift schedules.