Acute short-term sleep deprivation does not affect metacognitive monitoring captured by confidence ratings: a systematic literature review

This article presents the results of a systematic review of the literature surrounding the effects that acute sleep deprivation has on metacognitive monitoring. Metacognitive monitoring refers to the ability to accurately assess one's own performance and state of knowledge. The mechanism behind this assessment is captured by subjective feelings of confidence concerning the accuracy of our judgments or performance. These judgments influence decision behavior. How well these subjective feelings fit with reality is critical for good decision making. For example, a driver who is overconfident in their ability to remain vigilant after a night without sleep is at risk of having an accident. A learner who is overconfident in their ability to perform well on an exam without sleep is at risk of failing. A break down in metacognitive monitoring might be responsible for the increase in poor decision making observed when people are sleep deprived. Using defined search terms and exclusion criteria, electronic database searches identified ten empirical studies suitable for review. Participants in these studies completed performance-based tasks, typically cognitive, while remaining awake for 28-63 hours. In all studies, metacognitive monitoring was assessed via confidence ratings either pre-, on-, or post-task. Extended wakefulness had a significant negative effect on performance in most studies. Evidencing good monitoring, however, the monitoring estimates such as confidence also tended to decline. Moreover, two critical variables that assess the fit of these estimates to actual performance (bias and discrimination) were mostly unaffected by the number of hours awake. Still, some results indicated that these variables may be affected by substances intended to fight sleep deprivation, such as modafinil. Within the limitations of extant literature (e.g., a sampling bias towards young adult male participants), empirical observations to date converge to suggest that metacognitive monitoring remains largely unaffected by the examined quantities of acute sleep deprivation (up to 63 hours).