n spite of a century of scientific study of sleep, including three decades of modern intensive research, the function of sleep remains a biological enigma. This is not to say that there is a paucity of theories of sleep function. On the contrary, there are perhaps too many, given the relatively slim body of unambiguous data that can be marshalled in support of any one of them. Contrary to popular opinion, the function of sleep is not to prevent sleepiness, any more than the function of eating is to prevent hunger.
The term "function," when applied to a physiological or behavioral process, can take on different meanings depending on one's disciplinary interest and level of analysis. If we consider breathing (respiration using lungs) as an example, its function in oxygenating the blood is physiologically unambiguous. It is thought to have initially evolved in primordial vertebrates such as freshwater fish which were periodically exposed to drought. Thus, they could survive by breathing air when the rivers or lakes became dry. Breathing subsequently made possible the energetic requirements of terrestrial locomotion and flight through further evolutionary adaptations. One was utilized by higher vertebrates for vocalization in communication. Similarly, the possible functions of sleep can be considered first from a primordial evolutionary perspective, with the assumption that later modifications of a primordial sleep state function(s) could have evolved and served other adaptations or functions. Consideration here of the following hypotheses of the function(s) of sleep focus primarily on those which postulate a primordial biological function, because the evidence for such hypotheses is easier to evaluate than for those principally concerned with subsequent functions. In any case, the evidence pertaining to the various hypotheses indicates that no single hypothesis can account for all the data.
One hypothesis is that sleep serves to reverse and/or restore biochemical and/or physiological processes that are progressively degraded during prior wakefulness. This classical view of sleep function has prevailed over competing hypotheses, largely because it is so intuitively reasonable, and especially in light of the widespread detrimental psychological and behavioral effects that we all experience with a loss of sleep. Also consistent with this hypothesis is increased growth hormone secretion immediately following sleep onset in humans and baboons which remains synchronized with sleep even when normal sleep patterns are inverted from nighttime to daytime. This hypothesis is not supported by the finding of a decrease rather than increase in protein synthesis of the whole body during sleep in humans. The decrease in protein synthesis is attributed to sleep being a period of overnight fasting, since protein synthesis remains constant when subjects are fed continuously via intragastric tubes throughout the 24-hr period. Consistent with the restorative theory of sleep function are the increased amounts and "intensity" of sleep during a period of sleep recovery after 24 hours of sleep deprivation in humans and most other mammals.
Although it is often claimed that sleep deprivation conforms to the classical experimental paradigm of removing (i.e., extirpating) an organ to determine what derangements it produces, sleep is a physiological process involving practically all the organs of an animal. Furthermore, sleep itself cannot be removed without substituting something else in its place-notably wakefulness. Therefore, sleep deprivation does not strictly conform to the extirpation paradigm. Moreover, sleep deprivation cannot be performed without disrupting pre-established biological rhythms within the body. Studies of the effects of naps as well as sleep deprivation indicate that we tend to sleep longer the longer we have remain awake. But the results of studies on humans sleeping in constant environments without time cues, in which their biological clocks run freely without any influence from external cues, show that under these conditions sleep episodes tend to be shorter following long periods of wakefulness, which is contrary to the restorative hypothesis. On the other hand, under free-running conditions, the duration of sleep seems to be closely linked with the circadian rhythm of body temperature, so that the longest episodes are those that begin near the peak of body temperature (see Part G., Temporal Regulation of Sleep and Wakefulness). Studies on the effects of exercise on subsequent sleep also do not support the body restitution hypothesis. Increased sleep following elevated catabolism produced by physical exertion would be predicted, but the majority of more than 30 studies show no effects on post-exercise sleep. In those studies which did show an enhancement of NREM sleep, it can be concluded that the additional NREM sleep was a consequence of increased body temperature produced by the heating effects of exercise. Moreover, physically fit individuals do not have longer sleep durations or more NREM sleep than the unfit.
According to this hypothesis, sleep serves to reduce metabolic rate and body temperature in endothermic (warm-blooded) animals, mammals and birds during periods of rest to offset the high energetic costs of endothermy. The metabolic rate is 7-10 times greater in resting endotherms than in ectotherms (cold-blooded animals) when both are at the same body temperature. The strongest data in favor of the energy conservation hypothesis are the thermoregulatory and electrophysiological continuities between hibernation and sleep. Hibernation represents an extension of physiological processes initiated during sleep. When animals fall asleep their metabolic rates decrease by about 10% and heat is dissipated from their bodies through peripheral vasodilation, which leads to a 1-2°C reduction in body temperature. This reduction in body temperature is controlled through the reduction, at the onset of sleep, in the thermosensitivity of neurons of the peoptic nucleus of the hypothalamus below their prior waking threshold. This nucleus acts like a thermostat in the brain. When small mammals enter hibernation, they lower their body temperatures while remaining asleep, and continue to show predominant patterns of NREM sleep during hibernation. Because animals enter into hibernation from sleep, and in the early stages of hibernation there are NREM sleep patterns, and because hibernation has an energy conservation function, sleep may also reflect an energy conservation function.
However, the typical 10% reduction in metabolic rate below basal waking resting levels produced by sleep has been considered insufficient for the natural selection of sleep to have operated during the course of evolution. On the other hand, this 10% reduction in metabolic rate has typically been measured in animals exposed to thermoneutral temperatures. Energy savings can be considerably greater at lower ambient temperatures, since heat loss from the body increases proportionally as the temperature gradient between the body and the environment becomes greater. Moreover, in conditions of competition for available food, especially at low temperatures, a reduction of 10% in metabolic rate could make the difference between life and death for a small mammal or bird, as it has been calculated that a 100 g mammal exposed to freezing temperatures without food will survive a day longer if it dropped its temperature 2°C by sleeping, than if it remained resting quietly awake. The same forebrain structures and neurotransmitters that have been implicated in the control of sleep also seem to be involved in the control of body temperature.
Additional evidence in favor of the energy conservation hypothesis is the universal presence of NREM sleep in endothermic mammals and birds. On the other hand, although the typical characteristics of NREM sleep are not evident in reptiles and other ectotherms, they simply may not possess the requisite neural tissues necessary to produce EEG slow waves similar to that of NREM sleep in more encephalized endotherms. Heat production (thermogenesis) appears simultaneously with NREM sleep during infant development in altricial mammals, which is to be expected if NREM sleep serves to offset increased energy consumption required for endothermy.
The suppression or inhibition of motor activity during sleep prevents animals from attracting the attention of predators and, therefore, should decrease their vulnerability to predation. Although this hypothesis seems reasonable, it is difficult to evaluate what kinds of evidence would be crucial in its support or refutation. Moreover, it is also not clear why such a complex physiological process as sleep would have evolved merely to keep animals out of harm's way when waking behavioral inactivity would serve the same purpose. Many animals play "possum" by freezing while remaining awake when threatened by predators.
The inverse of this protective hypothesis is one that states that animals subject to predation are more vulnerable to predation while sleeping because of their decreased sensitivity to external stimuli during sleep, so that prey animals sleep less than predators. Of course this latter hypothesis is not a hypothesis about the function of sleep itself but rather about factors that may limit sleep. Nevertheless, this hypothesis is supported by negative correlations between amounts of both NREM and REM sleep and predatory danger. During REM sleep animals are even more difficult to arouse by stimulation than during NREM sleep. On the other hand, the rating of vulnerability to predation depends upon a variety of factors (e.g. degree of exposure, number of potential predators, potential defensive behaviors) and is, therefore, not easy to quantify reliably.
There have been numerous speculations on the functions of REM sleep, including stimulation of brain growth, fine tuning of the binocular oculomotor system, consolidation of memory, erasure of inappropriate memories, and the harmless discharge of strong emotions during sleep which would otherwise intrude into waking behavior. There is limited evidence favoring any one of these hypotheses, although the prolific amounts of REM sleep during fetal and infantile development seem especially compelling in arguments for a crucial ontogenetic role of REM sleep. One of these may be the maintenance of homeothermy of the brain during early stages when thermoregulatory mechanisms are incompletely developed because brain temperature rises during REM sleep.
Many of the foregoing theories stem from the apparent association of dreaming with REM sleep (See also Part H., Dreams). However, it should be acknowledged that although the higher probability of remembering a dream after awakening from REM sleep than from NREM sleep is consistent with the hypothesis that dreams occur most often during REM sleep, there is an alternate hypothesis, pertaining to the state-dependence of memory, that accounts equally well for differences in dream recall after REM and NREM awakenings. Clearly, more research is needed before firm conclusions can be drawn regarding either the functions of sleep as a whole or of its respective stages, but the fact that we all sleep and that there is such a persistent drive to attain sleep suggests that it subserves a vitally important function that enables us to remain alive on earth.
| Proceed to Part J. |
| Return to Syllabus Home. |