What is the role of sleep?

From supermemo.guru
Jump to: navigation, search

This text is part of: "Science of sleep" by Piotr Wozniak (2017)

Origins of sleep

Human brain is the highest achievement of the biological evolution. It all started from a simple ability to conduct impulses. Then the amazing concept of a neural network was developed. The brain of primitive vertebrates started adding new structures as well as new mechanisms for optimizing the jungle of neural connections. Sleep is a relatively old invention used to re-organize memories via molecular and neural mechanisms. Circadian rhythms are known in plants and in animals independent of the need for sleep. The process of evolution has, however, conveniently hooked up sleep to circadian rhythms to efficiently alternate between the explorative state (i.e. the use of the brain for learning new things) and the consolidation state (i.e. sleep). The circadian cycle has been associated with around a hundred known physiological functions and parameters that change in concert during the day (this number now increases rapidly, e.g. with circadian analysis of gene activation). Closely related to sleep are cycles in the levels of hormones such as serotonin and melatonin, ACTH and cortisol, acetylcholine, adenosine, and growth hormone. There is a circadian function that we can observe on our own without complex measurements: changes in the body temperature (see the figure Temperature changes in the course of the day in Biphasic nature of human sleep).

As we spend a third of our lives sleeping, there is little wonder that sleep has attracted lots of attention from neurophysiologists. Given the enormous complexity of the brain and its functions, there have been literally hundreds of theories that attempted to explain the role of sleep. Only recently, with the arrival of new research technologies, have we been able to see the big picture in the sea of detail.

Sleep theories

Over the last two centuries, dozens of theories of sleep have been proposed. Some scientists believe that we sleep to remember. Others believe that we sleep to forget. Yet others believe that sleep has nothing to do with memory. There are also theories that come from philosophers, religious figures, ideological movements, etc. For example, "sleep maximizes positivity" is a very vague "theory" that is actually largely correct.

For a biologist, the best ground for determining a validity of a theory is its evolutionary perspective. The main question to ask is: what vital function is subserved by sleep that demands turning off cognitive functions for a third of our lives! Each sleep theory must past this primary litmus test. Let's call it a shutdown test. Allan Rechtschaffen put it best saying "If sleep does not serve an absolutely vital function, then it is the biggest mistake the evolutionary process has ever made". The shutdown test is the best sieve for eliminating implausible theories of sleep, however, as an exercise, you can also check which theories explain the fact that if we do not get sleep on one day, we need more on another. For example, if sleep was to help avoid predation, we would not incur the predation avoidance debt that needs to be repaid. If we survive predation despite roaming around, our chances of survival on the following night are exactly the same.

We cannot forget, naturally, that some variables related to sleep do not need to be an expression of its primary function. In evolutionary terms, sleep is a very old phenomenon, and all species learned to attach dozens of neural and non-neural functions to this state. For example, sleep is the main anabolic state for the body. For this multifunction reason, it is hard to determine the primary function of sleep by just studying variables such as the sizes of animals, sizes of their brains, their habits, types of food, longevity, levels of activity, proportion of sleep spent in REM, etc. These all studies are very interesting but seem to explain little because of the huge number of overlapping variables which effectively obscure the main function of sleep.

Below I will quickly list theories that attempt to explain the role of sleep. I will divide them into those which answer and those which fail to answer that main question: why the conscious mind needs to be turned off. Let me start with a subset of countless theories that fail to pass the shutdown test (as suggested above).

Theories that fail the shutdown test

Here are just a few of innumerous theories that may be based on true facts, true models, or partially explain the function of sleep. However, they do not explain the essential need to "lose consciousness" for many hours each day:

  • energy conservation - If sleep was just a form of rest, why would we not just lie down in a dark place, while staying vigilant to predation. The brain is actually buzzing through the night as much as in waking, so turning it off would not offer any extra advantage. Sleep is no hibernation.
  • repair and restoration - If sleep was just to rebuild the body used up during the day, why does it not employ a simple metabolic reversal without a loss of consciousness? Many tissues can restore and repair in waking. Dedicated anabolic state is useful, but it could easily be enacted with the brain still on.
  • reactivation of infrequently used circuits - If activation of infrequently used memories was necessary to prevent forgetting, evolution might have simply employed molecular mechanisms in which once synapses were potentiated, they would spontaneously build up synaptic strength in time without secondary activation. Naturally, this would abolish the valuable evolutionary mechanism of forgetting. Similarly, random REM activation for the purpose of memory fixation would simply be wasteful. Forgetting plays a specific evolutionary function and the increase in synaptic strength should only result from a repeated exposure to a specific experience (for example, a repetition in SuperMemo serves the exactly this purpose)
  • recovery from slow-wave sleep - REM sleep was postulated as a recovery stage from NREM sleep. If REM was just a recovery stage, why should it happen several times during the night? Would it not just better come at the arrival of zeitgebers indicating the end of the sleep period? Why would we enter NREM sleep in the first place? NREM and REM clearly play a complementary function, but the theory says little about the need to enter the unconscious state of sleep.
  • avoidance of predation - Inactivity during sleep is supposed to minimize exposure to predators? However, why would we need turn off all the senses and extend reaction times to a near-useless degree? Moreover, carnivores tend to sleep far more than omnivores. This would hint that vigilance might be better than slumber in eluding the predator.
  • reversing oxidative stress damage - How does a loss of consciousness help the process of cleaning up free radicals? Carnivores sleep more than omnivores? Slow and big elephants need 3 hours of sleep while fast-living tiny rats may sleep for 18. Not all sleep variables need to be an expression of its neural function.
  • Siegel theory - Dr Jerome Siegel, a famous opponent of the link between sleep and memory, proposed his own theory of sleep: "sleep is merely a good way to increase an animal's efficiency, by hunting only when the hunting is good, for example, and minimize its risk" (Siegel 2009[1]). A tiny subset of the complex sleep control machinery would be sufficient to meet Dr Siegel's proposed function of sleep. For example, it would be enough to keep the animal awake and paralyzed as in REM sleep. All the rest of sleep control would be a simple waste of time and resources. For more on this theory see: Robert Vertes and Jerome Siegel
  • corneal anoxia - a theory says that REM sleep nourishes the corneas. Again, would not just waking up do the job if eyeball movement is all that is needed? If that function needs the anabolic state of sleep then it is only secondary to the main purpose of sleep. That eyes move in REM might simply be an expression of the fact that this is a movement that can do animal no harm. Possibly, eyes move just because their muscles do not need to be turned off.
  • promoting longevity - sleep is vital for survival, as such it contributes to longevity indeed. However, so does the liver, and no one claims that the main function of liver is to promote longevity. Moreover, the evolutionary process does not optimize for longevity, but for the perpetuation of a gene pool.
  • thermoregulation - some scientists believe that sleep serves thermoregulation while others attribute a large role for the body core temperature in generating sleep propensity. Dr Kurt Kräuchi summarizes his reasoning as follows: "all thermoregulatory effects which lead to an increase in the core/shell ratio (e.g. a reduced shell by increased distal skin temperature) lead to increased sleepiness and, as a consequence, to increased sleep propensity. However, the sleepiness/sleep regulatory system feeds back onto the thermoregulatory system only indirectly via sleep-related behaviors (e.g. relaxation, lying down)" (Kräuchi 2007[2]). That theory speaks a lot about causality, but very little about the purpose. It flunks the shutdown test.
  • detoxification - One of the fathers of sleep science, Romanian psychiatrist Constantin von Economo asserted that sleep was a detoxification process, in which the hypnotoxins were removed from the blood. However, some New Age authors who attempt to revive this view seem to commit a simple logical fallacy in which the brain shuts down for the purpose of preventing a hypnotoxic shutdown. The hypothesized function could well be done by the kidneys, the liver and other mechanisms that do not even come close to passing the shutdown test.
  • other theories: other theories of sleep such as psychoanalytical theories, Cayce theory, etc. border on paranormal and pseudoscience. They all may carry a grain of truth (e.g. neural network optimization will indeed serve self-edification), however, listing them all goes well beyond the scope of this article.

Theories that pass the shutdown test

If we consider the current status of knowledge about the function of sleep, many of the older theories start making sense. They all seem to converge into a central theme and all carry lots of inspiration. Here are some of these:

  • Krueger and Obal theory: the theory speaks of a natural propensity of neural networks to enter sleep states. However, the theory probably goes too far when calling sleep a network-emergent phenomenon that requires little or no central control (compare: Why we fall asleep?).
  • dreams create wisdom: yes! Dreams are an expression of a process that serves reorganization of memories. As a result of dreaming, knowledge becomes more coherent, more abstract, more applicable, etc. In sleep, skills get honed and fine-tuned. This means that we are wiser and more skillful in the morning than before going to sleep!
  • garbage collection: an early theory by Evans (1965), derived from computer garbage collection metaphor, says that sleep is needed to clean up the brain from useless memories. Indeed, there is a lot of evidence on synaptic weakening in sleep. Lots of garbage and chaos evaporates from the brain overnight.
  • reverse learning: inspired by the ideas of Evans, reverse learning was proposed in 1983 in Nature by Crick and Mitchison. Reverse learning theory, a neural equivalent of garbage collection, says that dreams sift away memories that are not needed, irrelevant or incongruent. Born in 1916, Francis H.C. Crick together with James Watson unraveled the mystery of the DNA double-helix for which they were awarded a Nobel Prize in 1963. In 1983, Crick and his colleague from Salk Institute, Graeme Mitchison argued in Nature (Crick and Mitchison 1983[3]) that human memory can get overloaded and REM sleep is used to run a garbage collection process on memories. They called this process reverse learning, and originally attributed it only one purpose: forgetting the unnecessary memory ballast. Crick named those ballast memories "parasitic memories", saying that "REM sleep allows us to have a smaller brain for a greater purpose". Crick's seminal 1983 article was taken further by many researchers who have ultimately concluded that REM sleep must serve optimization of memories expressed by synaptic weights of the neural network of the brain! If REM sleep was only to be used for forgetting the excess information, Crick's theories would run into the same evolutionary trouble as other theories listed above. After all, spontaneous loss of synaptic information with passing time would be a cheap solution that could be easily implemented metabolically. The value of the network optimization goes far beyond just forgetting.
  • consolidation of memory - some researchers posited that memories get consolidated in REM sleep. What is the purpose of NREM then? How can this be that antidepressants do not damage learning even if they are strong REM suppressants? Why does the REM brain behave like a waking brain? The term consolidation is a bit ambiguous. It may refer to processes occurring at the molecular as well as at the neural levels. Only neural consolidation might require turning off the brain. In reference to procedural learning a distinction has been made between acquisition, consolidation (post acquisition) and enhancement (sleep dependent). Of these, only the enhancement would depend on sleep. For a number of procedural tasks, sleep-dependent enhancement has been demonstrated (Stickgold and Walker 2007[4]). For more see: Neural optimization in sleep
  • psychotherapeutic theory - The theory says that in sleep we dream up situations that would be emotionally difficult to confront in real life. This is supposed to help us learn how to react in emergencies without actually living through those emergencies. This interpretation is plausible due to the fact that it involves an actual need to turn off the consciousness. Moreover, when dream contents are analyzed it is indeed found to be dominated by threatening situations. The psychotherapeutic theory, however, should rather be considered a narrow interpretation of a larger and more encompassing function of sleep: optimization of the neural circuitry. In this case, known information is played back in various constellations to compile new behavioral strategies that might be useful for survival. As I try to argue throughout this article, creative and optimizing function of sleep is more general and needed also for other neural functions, not only those related to the threats encountered in waking life.
  • activation and synthesis theory - Proposed by Hobson and McCarley in 1977 (Hobson and McCarley 1977[5]), Activation-Synthesis theory says that during sleep there is a sensory inhibition (input attenuation), and, during REM, there is a motor inhibition (output inhibition). Pontine PGO waves characteristic of the REM sleep travel up to the cortex that tries to interpret that stimulation resulting in dreaming. Visual, auditory and motor areas of the cortex are active and responsible for whatever we experience in dreams. Activation-Synthesis theory inspired further research into the neural function of REM sleep.
  • AIM model - the AIM model is a Hobson's improvement upon the activation-synthesis model that has dominated this field of research for two decades (Hobson et al. 2000[6])(picture). In the AIM model, brain states are described by 3 variables: (1) Activation, (2) Input source, and (3) Modulation. A healthy brain moves in selected trajectories in the AIM space. Various psychiatric and sleep disorders can be characterized as the departures from those trajectories. The model does not explain the function of sleep, but helps describe sleep in simple neurohormonal terms. For example, REM sleep is characterized by high activation, minimal sensory input, and cholinergic modulation.
  • brain development - In brain development theories, dreaming is supposed to provide internal stimulation of brain circuitry to ensure growth and development. As babies dream far more than adults, this theory finds inspiration in the correlation between the degree of immaturity in infancy and the proportion of REM sleep in animals. Human babies are particularly limited in their ability to interact with their environment and dreaming is said to act as a substitute for the waking experience. REM-induced muscle atonia combined with the brain activation would enable brain growth by activating the synapses without the negative consequences of acting out. The role of REM sleep in developing the visual system has been demonstrated by Stryker (Stryker et al. 2001[7]). REM deprivation results in developmental abnormalities later in life. Recently, the role of NREM sleep has also been confirmed.


  1. Siegel J.M., Sleep viewed as a state of adaptive inactivity, Nature Reviews Neuroscience / Volume 10 (October 2009): 747-753, doi: doi:10.1038/nrn2697
  2. Kräuchi K., "The human sleep–wake cycle reconsidered from a thermoregulatory point of view," Physiology & Behavior / Volume 90 / Issue 2-3 (February 28, 2007): 236–245, doi: 10.1016/j.physbeh.2006.09.005
  3. Crick F. and Mitchison G., "The function of dream sleep," Nature / Volume 304 (July 14, 1983): 111-114, doi: 10.1038/304111a0
  4. Stickgold R. and Walker M.P., "Sleep-Dependent Memory Consolidation and Reconsolidation," Sleep Medicine / Volume 8 / Issue 4 (June 2007): 331–343, doi: 10.1016/j.sleep.2007.03.011
  5. Hobson J. A., M.D. and McCarley R.W., M.D., "The brain as a dream state generator: an activation-synthesis hypothesis of the dream process," The American Journal of Psychiatry / Volume 134 / Issue 12 (December 1, 1977): 1335-1348
  6. Hobson J.A., Pace-Schott E.F., and Stickgold R., "Dreaming and the brain: Toward a cognitive neuroscience of conscious states," Behavioral and Brain Sciences / Volume 23 / Issue 6 (2000): 793-842
  7. Frank M.G., Issa N.P., and Stryker M.P., "Sleep Enhances Plasticity in the Developing Visual Cortex," Neuron / Volume 30 / Issue 1 (April 2001): 275-287