0202 Predicting Sleep Inertia in a Biomathematical Model of Fatigue and Performance: A Novel Approach. (25th May 2022)

• Record Type: Journal Article
• Authors: McCauley, Mark; McCauley, Peter; Kalachev, Leonid; van Dongen, Hans
• Is Part Of: Sleep. Volume 45(2022)Supplement 1
• Page Start: A92
• Page End: A93
• Publication Date: 2022-05-25
• Subjects: Sleep -- Physiological aspects -- Periodicals; Sleep disorders -- Periodicals; Sommeil -- Aspect physiologique -- Périodiques; Sommeil, Troubles du -- Périodiques; Sleep disorders; Sleep -- Physiological aspects; Sleep -- physiological aspects; Sleep Wake Disorders; Psychophysiology; Electronic journals; Periodicals; 616.8498
• DOI: 10.1093/sleep/zsac079.200
• Languages: English
• ISSNs: 0161-8105

Abstract

Abstract: Introduction: Biomathematical models of fatigue typically include sleep inertia as an additive process during wakefulness. However, there is predictive information to be gained from tracking the propensity for sleep inertia through sleep periods. We propose a novel approach involving a neurobiological sleep inertia process with relatively fast dynamics (in the order of several minutes) interacting with the much slower dynamics of the established processes of sleep/wake regulation. This sleep inertia process is captured by the addition of two ordinary differential equations (ODEs) in the model framework of McCauley and colleagues (2009, 2013, 2021) – one for wakefulness to track impairment from sleep inertia, and one for sleep to track the propensity for sleep inertia upon awakening. A single time constant is introduced to control the dynamic behavior of these ODEs to capture the dynamics of sleep inertia. Methods: 398 healthy young adults (ages 21–49 years) each participated in one of eight multi-day laboratory studies of total sleep deprivation, sustained sleep restriction, or simulated shift work. At 2–4 hour intervals while awake, participants performed a Psychomotor Vigilance Test (PVT), for which number of lapses (RT>500ms) was assessed, and rated their sleepiness on the Karolinska Sleepiness Scale (KSS). Sleep periods were recorded polysomnographically. Data were divided into a calibration set (five studies) used to estimate a single new model parameter capturing sleep inertia, and a validation set (three studies) used to independently verify model validity. Results: Based on the calibration data set, the sleep inertia time constant estimate was 0.71h±0.01. Based on the validation data set, goodness-of-fit root-mean-square-error was 2.28 for PVT and 0.733 for KSS, indicating high predictive accuracy. A dynamic buildup and then decline of predicted propensity for sleep inertia during sleep emerged, peaking 2–3h into the sleep period. Conclusion: The model expansion with a one-parameter sleep inertia process captured the transient effect of sleep inertia accurately across a range of sleep deprivation, sleep restriction, and simulated shift work scenarios. The emerging dynamic of sleep inertia propensity during sleep is consistent with findings on the magnitude of sleep inertia as a function of sleep duration and stage of awakening. Support (If Any): WSU HPC