Page:Risk of performance errors due to sleep loss, circadian desynchronization, fatigue, and work overload.pdf/9

Human Health and Performance Risks of Space Exploration Missions Cognitive impairments are present even after an individual has been awake for approximately 17 hours; in fact, recent studies have shown that these decrements are similar to those that result from an elevated blood alcohol level. A compelling Category I laboratory study from Williamson and Feyer (2000) used a cross-over randomized control design to observe cognitive and motor performance after minor sleep deprivation to performance after alcohol consumption. All subjects participated in both alcohol consumption and sleep deprivation, and the order of testing was counterbalanced so that half of the subjects participated in the alcohol consumption part first while the other half participated in the sleep deprivation part first. To avoid carry-over effects from one condition to the next, subjects were provided with a night of rest in a motel between each condition.

Results indicate that, on average, performance with a blood alcohol level of 0.05% remained equivalent to performance after being awake for 16.9 to 18.6 hours. Performance with a blood alcohol level of 0.1% was equivalent to performance after being awake for 17.7 to 19.7 hours, or to restricted sleep of 4 to 5 hours per night for 1 week (Czeisler, 2006). Similar studies that compare performance after a time of sleep deprivation to performance with elevated blood alcohol levels have confirmed these results (Dawson and Reid, 1997; Arnedt et al., 2001). These findings are compelling as the duration of wakefulness (17 hours), which results in decrements that are similar to those that are induced by a 0.05% blood alcohol level, is considered by many to be within the range of a "normal" waking "day"; many individuals can recall an incident in which they had to waken early in the morning and work all day and into the night. Astronauts, who sleep an average of 6 hours per night (Santy et al., 1988; Gundel et al., 1997; Monk et al., 1998; Kelly et al., 2005), may be performing critical tasks 17 hours or more after wakening.

Research suggests that circadian desynchronization and work overload may also impair performance. Specifically, a controlled laboratory study by Wright et al. (2002) evaluated the relationship between circadian rhythms and performance by assessing body temperature, which is regulated by the circadian mechanisms of the body. Body temperature is at its highest near the circadian peak and lowest near the circadian minimum (this is when the body is driven to sleep). It has long been recognized that a positive relationship exists between daily rhythms of body temperature and neurobehavioral performance and alertness in humans (Wright et al., 2002).

The study protocol (Wright et al., 2002) forced circadian desynchronization for 12 consecutive 28-hour days; participants were allowed 9.3 hours of scheduled time in bed and 18.7 hours of scheduled wakefulness. Performance on validated measures was evaluated every 2 hours, beginning 2 hours after the scheduled wake time. The protocol, therefore, assessed performance when the body is normally driven to sleep (which is related to the point at which body temperature at its lowest) relative to performance during normal waking hours, and allowed for assessment of the effects of body temperature independent of (and associated with) sleep hours and time of day. During the circadian peak (when body temperature is high), performance and alertness are high; conversely, near the circadian phase of low body temperature, performance and alertness are low. These results have been replicated in other forced desynchrony and extended wakefulness laboratory protocols (Wyatt et al., 1999).

Results from these laboratory protocols can be extrapolated to field conditions. Studies in the medical industry, where highly educated and trained individuals (e.g., physicians) are subject to circadian shifting and extended work shifts in addition to sleep loss, further demonstrate serious performance errors with populations that are analogous to astronauts. In a two-session, within-subject, Category II experiment that was conducted by Arnedt et al. (2005), the performance of 34 medical interns was observed under four conditions: