Fidelity of Information Processing on a Psychomotor Vigilance Task Predicts Changes in Self-Reported Sleepiness Ratings

INTRODUCTION: Daytime sleepiness is a common problem. Although sleepiness is primarily assessed at the self-report unit of analysis, factors that contribute to an individual’s experience and report of sleepiness remain poorly understood. While sleepiness is known to impact vigilance performance, the impact of vigilance performance on sleepiness reports is less well understood. We aimed to explore how performance on a psychomotor vigilance task (PVT) relates to changes in self-reported sleepiness in a rested condition. METHODS: Participants were 66 adults (Mdn=23, range 18–49 years old), 47% female, 88% white, with a wide range of insomnia symptoms. Participants rated their sleepiness on a scale from 1 (not sleepy) to 10 (extremely sleepy) at the start (pre) and the end (post) of a 10-minute computerized PVT. Ordinal regression determined whether mean reciprocal reaction time, a measure of overall performance, or the log-transformed signal-to-noise ratio (LSNR), a measure of fidelity of information processing, predicted post-sleepiness, adjusting for pre-sleepiness, insomnia, and potential confounds. RESULTS: Lower LSNR predicted greater change in sleepiness (pre-to-post PVT) and higher post-sleepiness even after adjusting for pre-sleepiness, mean reciprocal reaction time, insomnia, and other potential confounds (p<0.05). DISCUSSION: When adjusting for insomnia symptoms and potential confounds, participants with lower fidelity of information processing reported higher sleepiness than they had reported at the start of the PVT. Possible mechanisms and explanations are discussed in relation to a 3-factor model of sleep-wake states. This line of research may contribute to innovative approaches to assessing and treating sleepiness.