Sleep, time, and space—fatigue and performance deficits in pilots, commercial truck drivers, and astronauts

Abstract Sleep is essential for preventing fatigue in occupations that require sustained vigilance. We conducted a scoping review to synthesize knowledge about sleep, fatigue, and performance in pilots, commercial truck drivers, and astronauts. We found 28 studies where researchers objectively or subjectively measured sleep, fatigue, and performance. The research included laboratory-based (simulator) and field-based studies (i.e. real-world missions and a variety of shift-work schedules). Most researchers used actigraphy to measure sleep, and they found that ~6 hrs of sleep was common. The research also demonstrated how sleep duration and quality were negatively affected by schedule irregularity, early-morning start times, and high-risk missions (e.g. extravehicular activities in space). Collectively, the data demonstrated how shorter sleep durations, short off-duty time, and early-morning start times were associated with slower reaction times, more lapses in attention, and premature responses on psychomotor vigilance tests. Considering that few studies included polysomnography and circadian rhythm biomarkers, there remains limited knowledge about the effects of sleep microstructure and circadian rhythm alterations on performance abilities in these occupations. Future neurobiological and mechanistic discoveries will be important for enhancing vigilance, health, and safety for people working in the skies, on the roads, and in space. This paper is part of the David F. Dinges Festschrift Collection. This collection is sponsored by Pulsar Informatics and the Department of Psychiatry in the Perelman School of Medicine at the University of Pennsylvania.


Introduction
Sleep is critical for maintaining physical and psychological health. Adults require ~7 hrs of sleep each day, according to the American Academy of Sleep Medicine and the Sleep Research Society [1]. From an occupational health perspective, fatigue (a biological demand for recuperative rest) threatens safety and job performance [2]. Many highly publicized accidents have been attributed to fatigue, which underscores the importance of understanding how to promote healthy sleep and workplace/public safety [2]. Occupational responsibilities may restrict the times people have available to sleep. In the transportation industries, for example, people often work across multiple times zones, which can cause their circadian rhythms to become misaligned with their schedules [3]. In occupations that require nighttime work, people rely on their abilities to obtain enough daytime sleep, although this is opposed by humans' natural sleep-wake cycle patterns that favor wakefulness during the daytime and sleep during the night [4][5][6][7]. Work-related stress and environmental conditions (e.g. uncomfortable sleeping locations, sustained light exposure, and noise) also disrupt sleep and increase the risk for poor performance during work due to fatigue [2,8,9].
Airline pilots, commercial truck drivers, and astronauts are examples of occupations that require high-quality sleep for sustained vigilance during work. To perform their jobs effectively, they must endure stressful conditions and combine astute cognitive abilities with technical knowledge to make decisions safely under rapidly changing conditions [10][11][12][13][14][15][16][17][18][19]. Research has demonstrated how insufficient sleep impairs work-time functioning [12,20,21], which ultimately led to regulations that enforced duty restrictions and recommended fatigue-management strategies. Examples include Title 14 of the Code of Federal Regulations, which outlined duty limitations for safer aviation in 2014 [22]; the U.S. Federal Motor Carrier Safety Administration's 2014 revisions to duty time for truck drivers [23]; and the International Civil Aviation Organization's 2015 fatigue-management guidelines [24].
Safety-focused regulations were driven by the scientific discoveries of Dinges et al. [10,12,25,26], which established a battery of neurobehavioral performance metrics. Dinges and colleagues developed, tested, and validated methods for measuring vigilance and alertness, such as the 10-min Psychomotor Vigilance Test (PVT) [27,28] 3-min Brief PVT (PVT-B) [27,29,30], and Cognition Test Battery [31]. The PVT and PVT-B have been valuable for objectively quantifying responses to randomly timed visual stimuli to improve the rigor of occupational health research (compared with relying on methods that could be confounded by learning effects or subjectivity) [32,33]. Research participants' subjective descriptions of fatigue, for example, may not be consistently correlated with the objective measurements provided by the PVT.
In occupations where sleep times are restricted and there is a high-risk for catastrophic accidents, people are highly trained/ experienced, but they are not impervious to the adverse effects of attention lapses caused by inadequate sleep. Therefore, it is important for researchers to determine how sleep-related variables affect people's abilities for safe and effective job performance. Ultimately, the goal will be the development of optimal workplace practices that promote healthy sleep and mitigate the risks for errors and accidents. Previous reviews on this topic have focused on sleep deprivation rather than the occupationally relevant phenomena of short sleep and circadian misalignment [32,[34][35][36]. In addition, earlier reviews did not provide information about multiple occupations to allow for comparisons to understand the complex relationships among work schedules, sleep, fatigue, and performance in different settings. Therefore, the purpose of the present paper was to conduct a scoping review to synthesize research findings about objective and subjective measures of sleep, fatigue, and performance from three occupations-pilots, truck drivers, and astronauts.

Methods
A scoping review is a knowledge-synthesizing literature review that is guided by an exploratory question to determine key concepts, types of evidence, and research gaps [37,38]. We conducted our review according to Arksey and O'Malley's scoping review methodology [38]. The following question guided the review: What is known about sleep, fatigue, and performance in the safety-sensitive occupations involving aviation, truck driving, and space missions? We chose to examine multiple occupations to determine similarities and differences; this aspect of the scoping review was important for understanding whether findings were occupation-specific or common across professions.
To identify relevant research studies, we searched MEDLINE/ PubMed in August 2022 (excluding papers published before 2014 because several organizations implemented duty restrictions to increase sleep around this time) [22][23][24]39]. The keywords for the search were "sleep, circadian, fatigue, and performance" in combination with "airline pilots" (retrieved 12 citations), "truck drivers" (retrieved 4 citations), or "astronauts" (retrieved 5 citations). We screened the abstracts to exclude review papers and focus only on original research reports; ultimately, we selected 11 data-based papers about airline pilots (excluded 1 review), 3 about truck drivers (excluded 1 review), and 5 about astronauts. Additional articles were found by handsearching and examining article reference lists (leading to the addition of 6 additional papers about airline pilots and 3 papers about astronauts). A total of 28 research papers were used for the scoping review (Table 1).
The key findings across studies were organized to reflect the following themes: (1) sleep characteristics; (2) fatigue during work, (3) schedules and circadian rhythms; and (4) variables affecting performance.
The data from astronauts illustrated their difficulties with sleeping in space. Astronauts commonly used medications to sleep, such as zolpidem [43], zaleplon [43], and melatonin [55] (the use of sleep-inducing medications was not examined in pilots or truck drivers). According to polysomnography recordings analyzed by Koller et al. [55] , astronauts had a significant reduction in daily total sleep time by 0.8 ± 0.3 hrs when they were in space, compared with their sleep on Earth. In other studies about astronauts, shorter sleep durations and poor sleep quality were associated [54,59]; astronauts also rated the quality of their sleep in space poorly (compared with their sleep on Earth) [43]. Importantly, astronauts did not obtain >6 hrs of sleep on the nights before the more dangerous aspects of their missions, such as extravehicular activities (EVA) [43].
Koller et al. [55] hypothesized that microgravity may physiologically alter the structure of sleep in space. To test this hypothesis, four astronauts underwent polysomnography recordings. Koller et al. focused on two features of non-rapid eye movement (NREM) sleep: sleep spindles and slow-waves. Sleep spindlesbursts of oscillatory activity (9-15 Hz) lasting <2 s-have been associated with memory processing, while slow-waves indicate the depth and quality of NREM sleep. Sleep spindles can be fast (12)(13)(14)(15) or slow (9-12 Hz) and occur during stages 2 and 3 of NREM sleep. Compared with sleep on Earth, the astronauts' sleep in space was characterized by lower slow-wave amplitude (decreased by 2.5 ± 0.7 µV in-flight versus pre-flight, p = .01). They also demonstrated a higher fast spindle density (increased by 1.8 ± 0.5 spindles/min in-flight versus pre-flight, p < .001) and a shift toward higher frequencies in the slow spindles (increased by 0.2 ± 0.03 Hz in-flight versus pre-flight, which returned to baseline when astronauts returned to Earth). The following parameters were examined but did not differ significantly in the space/Earth comparisons: the duration of sleep stages, wake after sleep onset (WASO), fast spindle amplitude, slow spindle density, and slow spindle band-power. Collectively, the findings indicated that the depth of sleep may be reduced in space, and Koller et al. [55] also suggested that the sleep spindle changes could impact the ability to learn new skills, such as adapting to weightlessness in space, although this study did not demonstrate that any specific cognitive tasks were correlated with sleep microstructural changes .
The data from pilots illustrated how their fatigue levels were affected by numerous factors: rank/experience [40,67], in-flight nap opportunities [67], flight durations (short-haul [≤3 hrs] or long-haul [>6 hrs] flights) [65], take-off and landing times [42,49,64], high workloads (e.g. multiple assignments/landings per day) [40,52,53], and the duration of time spent awake [48,65]. A comparison of Captains and First Officers revealed that the more experienced pilots (Captains) reported lower Epworth Sleepiness Scale scores (7.7 ± 3.9 versus 9.7 ± 3.8, p < .001), and Fatigue Severity Scale scores were inversely correlated with pilots' age (r = −0.79, p < .0001) [40]. A qualitative analysis of pilots' perspectives about their napping opportunities on long flights supported the conclusion that experience provided them with skills for combating fatigue and staying focused-experienced pilots explained how they learned to adapt to their work conditions over time, which allowed them to use their in-flight/ pre-scheduled nap periods to effectively mitigate fatigue while their co-pilots were flying [67]. Long-haul pilots reported higher levels of fatigue (Samn-Perelli scale) when they had landings in the late evening or night [65], and fatigue scores were higher (Samn-Perelli and KSS at TOD) when pilots were flying between 0200 and 0600 hrs [62]. Two studies about short-haul pilots found that they did not experience significantly different Samn-Perelli scores at TOD depending on whether they flew earlier or later in the day [42,45], but pilots in another study had higher Samn-Perelli scale scores for early-morning versus late-morning flights (4.0 ± 0.9 versus 3.5 ± 0.8, p < .001) [42]. Data from the KSS (at TOD) also showed that pilots were significantly sleepier for early-morning versus nighttime landings (compared with midday measures) [49,62,64,65].
In pilots, there were inconsistent results about the effects of prior sleep duration on KSS and Samn-Perelli fatigue scale sores-some investigators found that pilots' prior sleep duration was significantly correlated with these fatigue measures [53,61] while others did not find significant associations between sleep duration and fatigue [65]. For example, Honn et al. [53] studied pilots who were using a flight simulator; when they had longer sleep before their simulated flights, their KSS and Samn-Perelli fatigue scores were significantly lower (p < .001)-every additional hour of sleep reduced their KSS scores by 0.4 ± 0.1 units and Samn-Perelli scores by 0.5 ± 0.1 units. In addition, Gander et al. [48] found that for every additional hour that short-haul pilots had been awake, there were significant increases in Samn-Perelli and KSS scores (at TOD) by 0.1 and 0.2 points, respectively.

Schedules and circadian rhythms
The reviewed studies involved many different types of work schedules. Unpredictable and irregular working hrs were particularly common in the transportation industry [40-42, 45, 48, 50, 52, 60-63, 65, 67]. Maritime pilots, for example, worked 'on-call' schedules that depended the movements of oil tankers, container ships, and cruise ships-these pilots' start times and duty durations were highly-variable [52]. Long-haul pilots had pre-determined schedules; however, they crossed multiple time zones during their flights, which shifted their clocks by 5-12 hrs [48,49]. Commercial airline pilots often flew multiple flights within a single day, and they had to remain alert for daytime, evening, and nighttime take-offs and landings [41,42,45,48,62,67]. Truck drivers had different types of schedules [50,56,63]. In the United States, for example, truck drivers were required to take a 34-hr break after accumulating 60-70 hrs of weekly driving. Although this requirement was intended to allow for rest, it led truck drivers' subsequent assignments to begin and end at different times [63]. For Australian truck drivers who worked in the coal mining industry, however, the work schedules were more predictable because they were assigned to day-or night-shifts (lasting 12-hrs), but their schedules rotated weekly [50].
The Morningness-Eveningness Questionnaire (MEQ; scores range 16-86) was used in four studies, which found that airline pilots and astronauts did not have strong dispositions for being awake or asleep at specific times of the day [44,45,58,66]. For example, the MEQ scores from astronauts [44], short-haul commercial pilots [45], and military pilots [58] indicated that they did not favor early or late sleeping and working times (e.g. MEQ scores of 52.9 ± 12.3 for astronauts, 51.4 ± 7.1 for commercial pilots, and 53.9 ± 8.2 for military pilots). The MEQ was not used in any of the studies about truck drivers [50,56,63]. In a stimulated space mission, MEQ scores were used to exclude subjects who reported definite preferences for morningness or eveningness (i.e. study participation required an MEQ score between 42 and 58 to avoid circadian preferences from confounding the study findings) [46].
Circadian biomarkers (e.g. urinary 6-sulfatoxymelatonin levels and 24-hr temperature fluctuations) were analyzed in airline pilots and astronauts [42,45,46]. For example, Arsintescu et al. [42,45] measured airline pilots' urinary 6-sulfatoxymelatonin rhythms to understand how flight schedules affected their circadian rhythms. They estimated a baseline acrophase (indicating the circadian nadir in each pilot's 6-sulfatoxymelatonin rhythm) as they flew for 5 consecutive days with mid-morning take-offs. Wide inter-individual circadian phase differences were found among the pilots during this baseline period-the urinary 6-sulfatoxymelatonin acrophases occurred between 0200 and 0630 hrs (n = 13). Then, after 3 to 4 days off, the pilots switched to a schedule requiring earlymorning take-offs, which was associated with a phase advance in most of the pilots (n = 9); however, two pilots had a phase delay-the sample's acrophases ranged between 0030 and 0450 hrs. When pilots' schedules rotated (to begin with midday take-offs), the sample demonstrated a mean phase advance of 1.3 hrs (n = 7 [phase advance]; n = 1 [phase delay]). When the pilots' shifts began in the evening, however, the 6-sulfatoxymelatonin rhythm acrophases occurred between 0200 and 0630 hrs. Although this small sample size limited statistical comparisons, these findings illustrated how pilots were able to acclimate to their schedules [42]. In a study of astronauts, Flynn-Evans et al. [46] tested the hypothesis that space travel would cause a misalignment between the endogenous circadian temperature rhythm and astronauts' sleep/wake schedules. The Circadian Performance Simulation Software was used to determine how astronauts' temperatures aligned with their sleep/wake cycles (considering that body temperature rhythms should typically reach a nadir during sleep). A percentage of the astronauts' temperature rhythms were not aligned with the timing of their sleep/wake periods 11 days before the mission (13%), in space (19%), and on the nights before conducting EVAs (29%). This temperature/sleep timing alignment was an important finding considering that astronauts with a circadian misalignment slept 0.5 fewer hrs and had VAS scores reflecting poorer subjective sleep quality (p ≤ .01) compared with astronauts without the misalignment [46]. The temperature/sleep timing and alignment was not significantly associated with astronauts' subjective reports of fatigue [46].
Signs of degraded performance on the PVT (e.g. slower reaction time [RT], more lapses in attention, and premature responses) were associated with having higher workloads (e.g. higher NASA Task Load Index subscale scores) [41], earlymorning start times [45,53], longer durations of wakefulness [53,66], shorter sleep before work [45,53], and elevated fatigue scores (on the KSS, Samn-Perelli scale, or VAS) [41,42,45,48,49,53,60,62,64,65]. Short commercial airline flights (which posed a higher workload for pilots according to the NASA Task Load Index) and early-morning take-offs (that reduced pilots' sleep duration) were associated poorer performance [41,45]. For example, comparisons of pilots rotating through schedules with different take-off times revealed how flying earlier in the morning was associated with a slower mean RT and more lapses in attention (257 ± [45]. In a flight simulator, pilots also demonstrated significantly more PVT lapses when they had to perform multiple take-offs and landings within a single day (as opposed to a single flight) [53]. The prior night's sleep duration was associated with the number of PVT lapses and premature responses-each additional hour of sleep reduced the false start rate by 0.4 ± 0.1 (p < .001) [53]. In a study of pilots who were flying between the U.S. and Japan, the duration of wakefulness and previous 24-hrs of sleep predicted the slowest 10% RT (at TOD). Every hour of wakefulness increased this PVT metric by 0.05 responses/s, while every additional hour of sleep was associated with an improvement by 0.10 responses/s [47].
The data for truck drivers indicated that off-duty time and rest periods, but not necessarily sleep duration, affected performance. Having >1 night to sleep between jobs was associated with significantly fewer lapses in attention on the PVT-B and fewer lane deviations while driving commercial trucks at night (despite no significant increase in their mean sleep duration during the off-duty period) [63]. A bio-mathematical model determined by Mollicone et al. used truck drivers' fatigue levels to predict hard-braking events. They determined that each lapse in the PVT increased the drivers' risk of hardbraking events by 8%, which demonstrates how poorer PVT performance can indicate an elevated risk for accidents [53]. Healthy volunteers also demonstrated performance impairments when they underwent simulated astronaut missions with sleep-restriction and deprivation [44,57,59,66]. When volunteers were sleep-deprived for 28 consecutive hrs before engaging in simulated spacecraft maneuvers, their PVT indicators worsened as they stayed awake (e.g. increase in lapses, slower RT) [66]. In the Human Exploration Research Analog habitat, volunteers' sleep was restricted to 5 hrs per day (for 5 consecutive nights, followed by 8 hrs of sleep for 2 nights), and their performance was significantly worse during the sleeprestriction period (e.g. slower RT). For astronauts, sleep durations <6 hrs on the ISS were associated with a significantly slower mean RT on the PVT-B [54], and ISS astronauts did not immediately return to their baseline level of performance post-mission. Astronauts required ~4 days post-mission to recover from deficits in motor functioning/perception as a result of space/microgravity (as determined by comparisons against healthy volunteers) [57].

Discussion
The present review synthesizes findings about sleep, fatigue, and performance in pilots, truck drivers, and astronauts-all three occupations require sustained periods of vigilance to prevent accidents, injuries, and deaths. Our scoping review found that sleep durations <7 hrs (the recommended sleep duration) [1] were common, particularly for truck drivers, astronauts (before and during missions), and pilots with early-morning takeoffs. The reviewed studies also demonstrated how longer sleep could significantly improve performance (likely translating into lower risks for errors, collisions, and accidents). Occupational schedules could cause circadian misalignment; therefore, sleep enhancement interventions, fatigue mitigation strategies, and the development of optimal work schedules will be important for pilots, truck drivers, and astronauts (key mechanisms are illustrated in Figure 1).
Upon recognizing how sleep significantly affects job performance, several governing bodies enforced regulations requiring specific break times for sleep in the trucking and aviation industries [22,39]. Despite the emphasis on breaks, however, the findings of present review illustrate how obtaining ≥7 hrs of sleep remains challenging. It is possible that occupational factors are impairing sleep, such as requirements for working rotating-shifts. Rotating schedules require periods of time when people must sleep during the day, but the schedules' rotating start and end times make acclimation difficult. Sunlight, an important circadian rhythm regulator, affects neural responses in the anterior hypothalamus, which regulates the pineal gland's melatonin production [7,68]. Melatonin levels typically begin rising when lights are dimmed, but shift-work alters the phase of melatonin rhythms. Consequently, shift-workers cannot always adapt their circadian rhythms to their work schedules, especially when schedules are unpredictable, which has been shown in police officers and nurses [17,69,70]. These findings emphasize the importance of considering the timing of sleep periods, not only the duration of off-duty time, when designing legislation about workplace health and safety. Complex tasks, learning, memory consolidation, and emotional regulation require adequate sleep. Arsintescu et al. [41] described how pilots' had higher workload complexities and demands with shorter flights-flight length and sleep duration were both negatively associated with the NASA Task Load Index sub-components, which measured effort, stress, and frustration. These findings illustrate the value of sleep for future endeavors related to workplace satisfaction, training, worker retention, and burnout prevention.
Astronauts' EVAs illustrate the highest-risk aspects of their work, considering the exposure to dangerous environmental conditions involving extreme cold and hypobaric hypoxia outside of the spacecraft [71,72]. Despite the dangers, the literature illustrated how astronauts have short sleep (~6 hrs), and they slept even less before EVAs. Considering how spaceexploration endeavors are likely to increase in regularity in the future, it will be imperative to develop protocols for ensuring that astronauts can obtain adequate high-quality sleep in space, despite the potential issues caused by weightlessness, changes in atmospheric gases, and the stressors associated with confinement.
Women comprise a minority within the aviation, trucking, and space-exploration workforces, and they were underrepresented in the reviewed studies (Table 1). Only one study specifically addressed the researchers' attempts to balance the numbers of men and women in their study design [59]. Considering that more women are likely to enter these occupations in the future, it will be important that studies are designed to have adequate statistical power for comparing sleep, fatigue, or performance measures by sex, especially considering previous reports of sex differences in circadian and homeostatic sleep-regulatory factors [73].
In conclusion, future research and policy efforts should focus on developing strategies to increase sleep duration and mitigate fatigue, in addition to advancing knowledge about sleep microstructure and circadian rhythms and their effects on performance abilities. Incorporating the neurobehavioral discoveries from Dinges and colleagues' [74] research programs into novel prediction models is important for prospectively determining the risks posed by various occupational activities (e.g. EVAs in space, ultra long-haul flights) according to sleep, circadian, genetic, and mission-related predictors to make well-informed decisions about performance and safety.