Estimated sleep from an under-mattress device predicts next-day vigilance, working memory, and mental arithmetic performance

Abstract Study Objectives Sleep is vitally important to maintain cognitive function, particularly in shift-work contexts. Sleep trackers can reliably estimate sleep, but the relationship between estimated sleep and specific cognitive domains are unclear. This study examined associations between estimated sleep and subsequent cognitive performance during a simulated night-shift protocol. Methods Twenty-four participants (meanSD age = 289 years) attended the sleep laboratory twice, for an 8-day simulated shift-work experimental protocol under two lighting conditions (standard- vs. circadian-informed lighting). Following a baseline sleep, participants remained awake for 27h and transitioned to sleeping between 10:00-19:00 with cognitive testing between 00:00-08:00 for four days. Tests included the Balloon Analogue Risk task, Continuous Performance task, Digit-Symbol Substitution test (DSST), Iowa Gambling task, Operation-Span task, Psychomotor Vigilance test (PVT), Stroop task, Tower of London task, and Trail Making test. Sleep was assessed using an under-mattress sensor (Withings Sleep Analyzer). Linear and non-linear models were used to test associations between estimated sleep and cognitive performance. Results Significant associations were found between sleep metrics and PVT reaction time (R2=.13, p=.008), Operation-Span arithmetic errors (.15, p=.023), proportion correct on DSST and Stroop tasks, (.06, p=.045; .09, p=.003), and Stroop reaction time (.19, p=.004). Random forest models demonstrated that vigilance, working memory, and mental arithmetic were associated with estimated sleep architecture, snoring, and cardiovascular function. Conclusions Sleep trackers could inform next-day cognitive performance, particularly vigilance, mental arithmetic, and working memory. In so doing, they may enable more informed interpretations of device-derived sleep and management of sleep-related cognitive impairment in future models. This paper is part of the Consumer Sleep Technology Collection.