LAUSR

The demands of a 24/7 society prevent millions of shift workers around the world from obtaining regular nightly sleep, negatively affecting their health, safety, and productivity. To mitigate this problem, workers should maximize the benefits of their atypical sleep opportunities. This requires assessing the ability of an individual to initiate and maintain sleep at the intended time, given their recent sleep history. Here, we present two validated mathematical models, one for predicting sleep latency and a second for predicting sleep duration, as decision aids for planning effective sleep schedules that maximize restorative sleep. As an extension of the two-process model of sleep regulation, the proposed models predict sleep latency and sleep duration by accounting for the nonlinear interactions between homeostatic sleep pressure and the circadian rhythm. We estimated the models’ parameter values for a population of young, healthy adults using experimental data from two studies and validated the models by comparing their predictions with experimental measurements from 22 different studies, involving a variety of sleep conditions spanning the entire circadian cycle and a total of 317 unique subjects. Comparison of the sleep-latency and sleep-duration models’ predictions with the experimental data showed that the models yielded acceptable average prediction errors: 4.0 minutes for sleep latency and 0.8 hours for sleep duration. Importantly, we identified conditions under which small shifts in bedtime may result in disproportionately large differences in sleep duration--knowledge that may be applied to improve productivity, safety, and sustainability in civilian and military operations. By extending the capabilities of existing predictive fatigue-management tools, the mathematical models described here allow users to determine the most opportune times to sleep. This information can be used to optimize sleep efficiency, allowing shift workers to maximize the benefits of restorative sleep.