Sleep Restriction, Sleep Deprivation, and Neurobehavioral Functioning

By Dr. Gregg D. Jacobs

A recent paper published in the journal Sleep by Dr. David Dinges and his colleagues at the University of Pennsylvania concerning the effects of sleep deprivation has drawn significant attention from the scientific community. The study, entitled "The Cumulative Effects of Additional Wakefulness: Dose Response Effects on Neurobehavioral Functions and Sleep Physiology from Chronic Sleep Restriction and Total Sleep Deprivation", examined the effects of different types of sleep deprivation on cognitive functioning.

Healthy young subjects with no psychological or physical health problems who averaged about 27 years of age and who usually slept about eight hours per night were assigned to one of four conditions: an eight hour sleep condition, a 6 hour sleep condition, a four hour sleep condition, and a no sleep condition (awake for 88 hours). Subjects in the eight, six, and four-hour sleep conditions were exposed to their condition for 14 consecutive days with no opportunity for recovery sleep or exposure to exercise or bright light. They were exposed to a variety of tests every two hours throughout the 14-day period that measured vigilance, performance, working memory performance, subjective sleepiness, mood and other symptoms.

The results of this study showed that, compared to subjects in the eight hour sleep condition, subjects in both the six and four hour sleep conditions exhibited impaired cognitive performance in a dose response fashion: the greater the number of days of sleep deprivation, the greater the deficits in cognitive performance. Additionally, cognitive deficits after two weeks of sleep deprivation in the four and six-hour conditions equaled those observed in the total sleep deprivation (88 hours awake) condition. However, subjective feelings of sleepiness did not follow the same pattern; in fact, subjects did not report increased sleepiness with greater sleep loss. For reasons that are unclear, the authors did not report the results of the mood measures (impaired mood is the most consistent effect of sleep loss) or the symptom data. (The fact that the authors did not report the results of these measures likely indicates that sleep loss did not have significant effects on these measures). The authors also found that subjects who habitually slept the longest in daily life prior to the study tended to exhibit the greatest impairment in cognitive functioning after the 14 days of sleep restriction.

Based on these findings, Dinges and his colleagues concluded that "since chronic restriction of sleep to six hours or less per night produced cognitive deficits equivalent to up to two nights of total sleep deprivation, it appears that even relatively moderate sleep restriction, if sustained night after night, can seriously impair waking neurobehavioral functioning in healthy young adults". However, when considering these findings and the authors' conclusions - and attempting to generalize their findings to sleep in daily life- several methodological issues about this study must be kept in mind.

First, many studies are affected by what researchers call "demand characteristics", which is an attempt by subjects to satisfy the experimenters' expectations about the outcome of the study. Although the authors of this study did not describe how they explained the rationale for the study to subjects, it is very difficult for experimenters who are noted sleep deprivation researchers not to convey, either consciously or unconsciously, their expectations to subjects (e.g., that sleep loss will have adverse effects on cognitive performance). Subjects who are influenced by these expectations can exhibit behaviors that are consistent with the experimenters' expectations in an attempt to "satisfy" the experimenters. This problem is compounded by the fact that many subject's expectations are influenced by media reports that emphasize the adverse effects of sleep deprivation.

Second, other studies that have assessed the effects of sleep loss have not found the same adverse effects as those reported in the Dinges study. Although Dinges and his colleagues suggest that this was due to the fact that prior studies employed insensitive tests, lacked sophisticated data analysis techniques, employed poor control over actual sleep schedules, and allowed use of caffeine, it is also just as likely that the effects of sleep loss are inconsistent and not very robust; otherwise, they would be consistent across different studies, outcome measures, and data analysis techniques.

Third, this study employed a complex, obscure data analytic strategy that is difficult to interpret and does not allow comparisons to other studies that use more typical analysis of variance approaches for data analysis. This study would have been strengthened by the use of both techniques to determine whether more traditional data analysis techniques yielded similar results. If the effects of sleep loss are robust and consistent, they should be observable with more traditional data analysis techniques.

Fourth, the results of this study can only be generalized to young, healthy eight-hour sleepers, which excludes a significant number of adults ages 30 and older who are shorter sleepers. Recent polls by the National Sleep Foundation and a major epidemiological study that assessed sleep durations in a population study of over a million people found that the majority of Americans sleep less than eight hours per night, which means that the subjects in the Dinges study may be an atypical group of longer, younger sleepers. It may be that young people who sleep eight hours per night are more affected by chronic sleep loss than other age groups or shorter sleepers. Indeed, the authors found that the greatest decrements in cognitive performance were, in fact, observed in the longer sleepers.

Fifth, subjects were not subjectively sleepy and no data was presented to indicate that sleep loss produced effects on mood or other symptoms. For the majority of people in daily life, the most distressing effects of sleep loss are impaired mood and subjective alertness, yet these were not compromised even under the rather extreme conditions of sleep restriction employed in this study.

Sixth, the neurobehavioral measures in this study were not significantly compromised until after a week without any recovery sleep. This type of sleep loss is not representative of sleep in daily life since most people alternate between sleep loss and recovery sleep. (Similarly, it is very atypical for people to have no opportunity for exposure to physical activity or bright light for two weeks, both of which produce alertness effects that counteract the effects of sleep loss.) Indeed, the 2002 National Sleep Foundation poll found that Americans average just under seven hours of sleep during the week but increase their sleep duration on weekends. This suggests that the majority of adults do not experience sleep restriction for two weeks without recovery sleep as subjects did in this study. Furthermore, the individuals in the six and four hour sleep conditions actually slept about 5.5 and 3.75 hours, respectively, for two weeks without recovery sleep. This means that the authors' conclusion that "sleep restriction to six hours or less per night can seriously impair neurobehavioral functioning" is not really accurate; it also means that subjects' baseline sleep duration in the six and four hour sleep condition was restricted by over 30% and 50%, respectively, for two weeks. (Longer sleepers, who exhibited the greatest impairments in neurobehavioral functions from the 14 days of imposed sleep restriction, likely experienced an even greater percentage reduction of their baseline sleep duration). Chronic reduction of this percentage of sleep for two weeks, without recovery sleep, is not likely to occur in daily life.

In any scientific study, it is important to consider the findings of the study in the context of the types of subjects used, the conditions and measures employed, and the conclusions reached by the researchers. Although the Dinges study suggests that sleep loss can have significant effects on some aspects of daytime functioning (cognitive performance), these effects may only occur for eight hour sleepers whose sleep duration is reduced significantly and for extended periods of time without recovery sleep. Other studies employing more realistic assessments of the effects of the types of sleep loss experienced by most people in daily have shown that the effect of sleep loss are inconsistent and subtle. In fact, insomniacs (who represent a significant proportion of the population and who average about six hours of sleep) do not exhibit decrements in cognitive functioning.

If we are to gain a more complete understanding of the effects of sleep deprivation, issues such as what kind of sleep loss (mild, moderate, or significant reduction in baseline sleep duration), duration of sleep loss (days or weeks) with or without recovery sleep, and sleep loss in whom (long or short sleepers, young or old) must be addressed in future research. Future studies should also assess the effects of other sleep-related behaviors on neurobehavioral functioning that are common in daily life and of equal or greater concern than the type of sleep deprivation examined in the Dinges study. For example, a significant portion of the population uses sleeping pills on a chronic basis. Although sleeping pills are known to compromise psychomotor functioning and have been linked to increased morbidity, few well-controlled studies have assesses the neurobehavioral effects of their chronic use. With more people likely to take sleeping pills due to the widely disseminated recommendation by some sleep researchers that we need at least eight hours of sleep per night, such a study would be timely and important.