Sleep as Biology

1. The Question That Is Rarely Asked

When sleep researchers discuss what sleep does, the conversation moves quickly to familiar territory: memory consolidation, immune function, cellular repair, hormonal regulation. These are real processes. The research supporting them is credible and substantial. But the question that is less often asked — and that yields a more illuminating answer — is what sleep is. Not what it does during the hours it occupies, but why those hours exist at all. Why do animals become unconscious, vulnerable, and unresponsive to their environment at regular intervals? What is the evolutionary logic that made this dangerous-seeming state worth keeping?

The dominant popular assumption — that sleep is primarily a restorative process, a period of biological repair — answers the what but sidesteps the why. If the body required repair, evolution had millions of years and considerable ingenuity to find other means of achieving it. Many biological maintenance processes occur continuously and do not require the complete suspension of environmental awareness that sleep entails. The restoring-while-resting account leaves the core puzzle intact.

Jerome Siegel, a sleep researcher at the University of California, Los Angeles, proposed a more fundamental answer in 2009. Sleep, he argued, is best understood not as a restorative state but as a state of adaptive inactivity — a period during which an organism withdraws from metabolic expenditure when activity would offer no survival advantage, or when the costs of remaining active would outweigh any benefit (Siegel, 2009). Rest, on this account, is not a consequence of exhaustion; it is a strategy. The body is not simply switched off while the brain repairs itself. It is actively choosing the most energy-efficient posture in relation to its environment.

This distinction matters. If sleep is understood primarily as restoration, then any disruption to it becomes a threat to the restoration process, and the calculus of harm tilts heavily toward pathology. But if sleep is understood primarily as adaptive inactivity — as a flexible biological strategy calibrated to the organism's environment and needs — then the variation observed across species, across life stages, and across individuals becomes not a problem to be solved but evidence of the strategy working as designed. The animal kingdom makes this case with considerable force.

2. What Animals Do

2.1 The Elephant's Two Hours

African elephants in the wild sleep for approximately two hours per day (Gravett et al., 2017). In captivity, where there are no predators to monitor, no distances to travel, and food is reliably provided, that figure rises considerably. The same species, in different environmental conditions, exercises a profoundly different sleep pattern. The wild elephant's two-hour schedule is not a deficit — it is a calibration. When extended vigilance is necessary, the system allocates accordingly. When it is not, the system relaxes. What changes is not the biology of sleep; it is the balance of advantage between wakefulness and rest in that particular environment.

This finding is not idiosyncratic to elephants. Across large mammals studied in natural habitats, sleep durations are consistently shorter than those recorded in captivity, and they vary in response to ecological conditions — predation pressure, foraging demands, seasonal changes in food availability, reproductive state (Siegel, 2009). The body does not have a fixed sleep requirement, like a battery that must reach a specific charge level. It has a system that weighs the costs of wakefulness against the costs of inactivity and adjusts accordingly.

2.2 The Swift's Aerial Sleep

The common swift (Apus apus) spends the majority of its life airborne. It feeds in flight, drinks in flight, mates in flight, and — critically for the argument being developed here — sleeps in flight. Electrophysiological research on great frigatebirds, a species with comparable flight patterns, has demonstrated that birds can sleep with either one cerebral hemisphere at a time or both hemispheres simultaneously during non-stop oceanic flights lasting up to ten days (Rattenborg et al., 2016). During those flights, the birds slept for an average of 0.69 hours per day — a fraction of the twelve hours or more spent sleeping on land. Yet they remained functionally competent throughout.

The swifts of Bungay — the birds that return each May to the eaves of Nethergate Street and disappear each August back towards Africa — are celebrated in a companion essay on this platform for the remarkable biology that keeps them airborne for ten months of the year (see: Swift Living, Swift Work). Their sleep architecture is part of that story. What the frigatebird research makes explicit is the mechanism: a brain that can satisfy the requirements of sleep hemispherically, keeping one half in a state sufficient to maintain flight control while the other rests. The swift's body is not cheating sleep; it is expressing sleep in the form that its ecology demands.

This is the adaptive inactivity hypothesis made visible. Sleep is not a binary — on or off, present or absent — with a fixed daily quota. It is a state that the organism calibrates to its circumstances, taking what it can when it can, in the form that is compatible with survival.

2.3 Unihemispheric Sleep and the Dolphin

The bottlenose dolphin presents a version of the same solution to a different ecological problem. As a mammal, it must breathe air; as a marine animal, it cannot afford to lose consciousness in water. The solution is unihemispheric slow-wave sleep: one cerebral hemisphere sleeps while the other remains active, enabling the animal to surface and breathe, maintain motor coordination, and continue monitoring its environment (Rattenborg, Amlaner, and Lima, 2000; Lyamin et al., 2008). After a period, the hemispheres switch. The dolphin is never fully unconscious, yet it achieves the neural restoration associated with sleep.

This is an architectural solution to the sleep-survival tension, and it is not unique to dolphins. Unihemispheric sleep has been recorded in a range of species, including fur seals, manatees, and multiple avian species, with the open eye directed toward perceived threats and away from group members (Rattenborg et al., 1999). In mallard ducks sleeping at the exposed edge of a group, the proportion of unihemispheric sleep increases dramatically compared to those in the protected centre — the system responds to predation risk in real time (Rattenborg et al., 1999). These are not aberrations. They are evidence of a biological system so flexible that it can modulate not only its duration but its very architecture in response to environmental demand.

2.4 What the Animal Evidence Shows

Taken together, the evidence from elephants, swifts, frigatebirds, dolphins, and mallards does not describe a single, universal sleep requirement that all organisms express in different ways. It describes a conserved biological principle — adaptive inactivity — that different organisms have implemented through radically different mechanisms, durations, and architectures. The common thread is not a fixed quantity of rest. It is a system that withdraws from the costs of wakefulness when those costs cannot be justified by the benefits, and does so in whatever form is compatible with the organism's particular survival constraints.

The implications for understanding human sleep — and human infant sleep in particular — are significant, and will be examined in the sections that follow.

3. The Mechanism: What Happens During Sleep

Understanding sleep as adaptive inactivity does not require dismissing the biological processes that occur during it. On the contrary, those processes are most coherently understood as expressions of the energy conservation strategy, not as the primary purpose that explains why sleep exists.

3.1 Metabolic Suppression

During sleep, body temperature drops, heart rate slows, digestive activity reduces, and overall metabolic rate falls measurably below waking levels. In humans, this metabolic reduction is estimated at approximately 15–25% compared to resting wakefulness, with the suppression deepest during slow-wave sleep (Horne, 2012). This is the energy conservation mechanism working directly: the body is operating on a reduced budget, diverting resources away from the costly maintenance of full alertness and physical readiness.

In animals facing seasonal or environmental extremes, this same metabolic suppression logic is taken to remarkable lengths. The African lungfish and the polar bear, examined in the companion essay in this series (see: Sleep Across the Spectrum), represent the continuum from human sleep to full seasonal dormancy — and make visible the underlying principle that sleep, torpor, and hibernation are expressions of the same adaptive logic rather than categorically separate states. The current essay focuses on the human end of that continuum; the broader biological spectrum is examined there.

3.2 Neural Processes

During sleep — particularly slow-wave sleep — the brain undergoes a process now described as glymphatic clearance: the cerebrospinal fluid system becomes dramatically more active, flushing metabolic waste products including amyloid beta, the protein associated with Alzheimer's pathology (Xie et al., 2013). This finding has attracted considerable attention, not least because it appears to provide a cellular mechanism for the correlation between chronic sleep disruption and neurodegenerative risk. It is real and important neuroscience.

The point to be made here, however, is that glymphatic clearance does not require sleep per se — it requires the low neural activity and altered cerebrospinal dynamics that characterise sleep. Sleep is the context in which these conditions are most reliably achieved; it is not the process itself. This is a distinction with consequences. It suggests that it is the metabolic state of low activity — the energy conservation state — that enables these neural maintenance processes, not the reverse. The brain cleans itself during sleep because sleep is when the neural traffic quietens sufficiently for the cleaning system to operate effectively. The cleaning is facilitated by the inactivity; it is not the reason the inactivity evolved.

3.3 Growth Hormone and a Cautionary Note

One of the most persistent claims in popular accounts of sleep is that growth hormone is released during sleep, and therefore that sleep deprivation stunts children's growth. The association between slow-wave sleep and growth hormone secretion is well established and dates to the classic studies of Takahashi, Kipnis, and Daughaday (1968), who demonstrated that in men, approximately 70% of growth hormone pulses coincide with slow-wave sleep episodes. This association is real and has been replicated.

The causal mechanism, however, is considerably less certain than popular accounts suggest. A 2022 study by Shaw and colleagues, published in the Journal of the Endocrine Society, disrupted slow-wave sleep in pubertal children by 40% using auditory stimuli across an overnight study. The result was unexpected: there was no diminishment of basal or pulsatile growth hormone secretion in either boys or girls (Shaw et al., 2022). The researchers' interpretation is significant: slow-wave sleep itself may not directly stimulate growth hormone secretion. Rather, a common upstream neural signal — most likely growth hormone-releasing hormone (GHRH) — may simultaneously induce slow-wave sleep and growth hormone secretion. Disrupting the sleep stage does not interrupt the upstream signal.

A 2023 narrative review examining the relationship between growth hormone and sleep in children characterised the available evidence as "limited, old, poor quality, and heterogeneous and inconsistent" (Zaffanello et al., 2023). This is not a fringe critique. It is the considered assessment of the evidence base by researchers working within the field.

The growth hormone/sleep story is a precise illustration of the broader pattern this essay is tracing: the association is real; the causal mechanism claimed by popular science is not robustly supported; and the anxiety generated around disrupted sleep — including in parents of infants — is disproportionate to what the evidence shows. The child who wakes three times in the night is not being deprived of growth hormone. The upstream signal for that hormone's secretion appears to operate independently of the sleep stage with which it has long been associated.

4. Polyphasic Sleep: Flexibility as the Default

The assumption that healthy human sleep consists of a single consolidated overnight block is so culturally embedded that its relative recency as a norm is rarely acknowledged in clinical or popular contexts. The historical construction of that norm — its origins in the industrial transformation of working hours, the imposition of the factory clock, and the extension of artificial light into the night hours — is examined in the third essay in this series (see: Sleep as Culture). The current discussion concerns the biology underlying the historical pattern.

Many animals, including humans, are naturally inclined toward polyphasic sleep — multiple sleep episodes distributed across the twenty-four hour cycle. The post-lunch dip in alertness, experienced across cultures and time zones, reflects a biological oscillation in the circadian drive for sleep that is independent of meal timing and appears to be a genuine feature of human physiology (Dinges, 1992). In cultures where the afternoon rest is socially permitted — the Mediterranean siesta, the tropical midday rest — the biological inclination finds expression. In cultures where it does not, the inclination remains but is suppressed by social and occupational demand.

Infants sleep polyphasically because infant biology is polyphasic. The consolidated overnight sleep that parenting culture treats as the goal — and the infant's failure to achieve it as the problem — is not a biological standard being failed. It is a culturally constructed preference that is biologically unreasonable for an infant to meet. The adaptive inactivity framework explains why: infant sleep is calibrated to infant metabolic needs, feeding requirements, and developmental demands. A newborn brain is not a small adult brain operating on reduced power; it is a radically different system, running intensive developmental processes that require frequent interruption and feeding. Multiple short sleep episodes are not a malfunction of the infant sleep system. They are the infant sleep system operating as designed.

This has direct implications for parental wellbeing that are explored further in the HWTK companion piece Why Don't Babies Sleep Through the Night — and Why That's Not a Problem and in the plain-language IOW version of this essay (see: IOW — Where the Idea of Eight Hours Sleep Actually Came From).

5. The Same Signal in a Different Register

The adaptive inactivity principle operates not only in response to external environmental conditions — darkness, predation pressure, the absence of food — but also in response to internal physiological states that share the same underlying logic. Depression and grief provide important examples.

The lethargy associated with major depressive disorder, and the exhausted withdrawal of acute grief, are routinely framed as symptoms to be overcome — departures from normal functioning that represent the illness rather than any form of adaptive response. The neuroscientific literature on depression suggests a more nuanced picture. The cytokine hypothesis of depression, developed by Dantzer and colleagues among others, proposes that the behavioural features of depression — reduced activity, social withdrawal, decreased appetite, hypersomnia — closely resemble the sickness behaviour observed across mammals in response to immune activation (Dantzer et al., 2008). When the body is fighting an infection, it actively reduces its external activity to redirect metabolic resources toward the immune response. The lethargy is not incidental to the immune fight. It is part of it.

Depression, on this account, may represent a similar adaptive withdrawal in response to a different class of signal — psychological threat, unresolvable stress, loss — in which the body's ancient energy-conservation machinery is activated in the absence of the infectious trigger that originally shaped it. The withdrawal is intelligible as biology, even when the circumstances that have produced it are psychological rather than immunological.

This does not mean that depressive withdrawal should be left unchallenged, or that grief requires no support. It means that the withdrawal itself is intelligible as biology, and that treating it solely as dysfunction — as something to be corrected and overridden — misses the adaptive logic operating beneath it. The parallel with physical healing is precise: one does not treat a fracture by insisting the patient resume full activity immediately, because the withdrawal from activity is part of the repair process. The appropriate response is to support the conditions under which the system can do what it is attempting to do, whilst providing whatever external scaffolding prevents the withdrawal from becoming so complete that it forecloses recovery.

The full argument is developed in Natural Healing: Understanding Recovery Across Physical, Psychological, and Therapeutic Domains, which this essay treats as a companion piece. The point relevant here is this: the lethargy of the clinically depressed individual and the exhaustion of the newly bereaved are not departures from the adaptive inactivity principle. They are expressions of it, operating in response to a different class of environmental signal — not darkness, not predation pressure, not the absence of food, but the internal equivalent: a system under load that cannot be resolved by further activity.

6. Freud and the Defended State

A brief and perhaps unexpected witness for the prosecution — or, more accurately, for a convergent intuition — is Sigmund Freud. Freud's name in the context of sleep is most readily associated with wish-fulfilment: the theory that dreams are the disguised expression of repressed unconscious desires (The Interpretation of Dreams, 1900). This is the popular headline, and it is the aspect of Freud's theory that has fared worst under empirical scrutiny. The activation-synthesis hypothesis proposed by Hobson and McCarley (1977) offers a more parsimonious neuroscientific account: dreams are the cortex making narrative sense of random brainstem activation during REM sleep, not the product of an elaborate unconscious censorship mechanism. Freud's interpretive machinery — condensation, displacement, latent and manifest content — finds little support in the evidence.

What is less often noted is that Freud's account of sleep itself, as distinct from his account of dreams, contains an insight that sits in surprising alignment with the adaptive inactivity thesis. Freud argued that dreams function as guardians of sleep — the unconscious generates narrative to absorb stimuli that would otherwise wake the sleeper, preserving the sleeping state as a biological priority. Sleep, in Freud's framework, is the primary state being defended. The dream is the mechanism of its defence. Freud arrived at this position through a theoretical apparatus that is not empirically credible; but the functional claim — that the organism actively works to preserve and defend the sleeping state — has an echo in the energy conservation account. The body that reconfigures its neural architecture hemispherically to sleep while flying, or that calibrates its sleep duration to predation pressure, is a body for which sleep is sufficiently valuable to be worth defending by sophisticated means. Freud intuited a biological priority. He simply had the wrong mechanism.

7. The Anxiety That the Evidence Does Not Support

The preceding sections have established the following: that sleep is a flexible, adaptive biological system calibrated to environmental demands; that its duration, architecture, and distribution across the day vary substantially across and within species; that the association between sleep and specific biological processes — including growth hormone secretion — does not reliably support the causal claims made in popular accounts; and that polyphasic sleep, including infant polyphasic sleep, is biologically coherent rather than pathological.

None of this is to say that sleep is unimportant. The body does not maintain complex systems without good reason, and the evidence that severe or chronic sleep deprivation has negative health consequences is credible. The argument being made here is more precise, and more defensible, than the claim that sleep does not matter. It is this: the evidence that variation from culturally constructed norms — sleeping less than eight hours, waking during the night, sleeping in episodes rather than a consolidated block — constitutes a medical or developmental emergency is weak. The anxiety layered on top of the genuine biology of tiredness is disproportionate to what the research actually shows.

The sleep anxiety industry — the sleep charities, the public health messaging, the popular science books that frame insufficient sleep as a civilisational catastrophe — has a legitimate foundation. Extreme sleep disruption is harmful. But the application of that legitimate concern to the normal variation of human sleep, and in particular to the normal sleep behaviour of human infants and the parents who care for them, extends well beyond what the evidence supports. The parent at three in the morning, feeding a waking infant for the fourth time, is not watching their child's development derail. They are participating in a process that evolution has been running successfully for the entirety of human prehistory.

The cultural construction of that anxiety, and the historical process that produced the norms against which normal human sleep is now judged pathological, is the subject of the companion essay Sleep as Culture. The biological spectrum that extends from human sleep through torpor and aestivation to full seasonal dormancy — and what that spectrum reveals about the underlying adaptive principle — is examined in Sleep Across the Spectrum.

8. Conclusion: What Sleep Actually Is

Sleep is energy conservation. It is the biological expression of a principle that runs from the human infant's fragmented night, through the elephant's two alert hours on the savannah, to the frigatebird sleeping on the wing above a moonlit ocean. The organism withdraws from the metabolic cost of wakefulness when the environment offers no advantage sufficient to justify that cost, in whatever form and for whatever duration its particular constraints allow.

The restorative processes that occur during sleep — the glymphatic clearance, the memory consolidation, the hormonal pulsing — are real, and they are important. But they are consequences of the low-activity state that energy conservation produces, not the primary explanation for why that state evolved. Sleep did not evolve in order to consolidate memories. It evolved because staying awake in the dark, without predators to avoid or food to find, costs energy that cannot be recovered. The biological machinery that maintains the sleeping organism took advantage of the quietened state to perform maintenance tasks; it did not create the sleeping state in order to perform them.

This matters because it changes what we are looking for when we ask whether someone is sleeping well. The question is not whether they are meeting a culturally specified quota. The question is whether their biological system is calibrating appropriately to their circumstances — withdrawing when withdrawal is advantageous, rousing when attention is required, distributing rest across the periods when activity offers least return. By those criteria, the infant waking to feed is calibrating perfectly. The parent exhausted by those wakings is bearing a genuine cost. The anxiety they are made to feel about whether something is wrong — with their child, with their parenting, with their family's sleep — is, to a very considerable extent, the sleep industry's product.

The biology, stripped of its cultural scaffolding, offers a different and more generous account.

---

References

Dantzer, R., O'Connor, J.C., Freund, G.G., Johnson, R.W. and Kelley, K.W. (2008). 'From inflammation to sickness and depression: when the immune system subjugates the brain.' Nature Reviews Neuroscience, 9(1), 46–56.

Dinges, D.F. (1992). 'Probing the limits of functional capability: the effects of sleep loss on short-duration tasks.' In R.J. Broughton and R.D. Ogilvie (eds), Sleep, Arousal, and Performance. Boston: Birkhäuser.

Freud, S. (1900). The Interpretation of Dreams. Vienna: Franz Deuticke. [Standard Edition translation, Hogarth Press, 1953.]

Gravett, N., Bhagwandin, A., Sutcliffe, R., Landen, K., Chase, M.J., Lyamin, O.I., Siegel, J.M. and Manger, P.R. (2017). 'Inactivity/sleep in two wild free-roaming African elephant matriarchs — does large body size make elephants the shortest mammalian sleepers?' PLOS ONE, 12(3), e0171903.

Hobson, J.A. and McCarley, R.W. (1977). 'The brain as a dream state generator: an activation-synthesis hypothesis of the dream process.' American Journal of Psychiatry, 134(12), 1335–1348.

Horne, J. (2012). 'Working throughout the night: beyond "sleep homeostasis".' Neuroscience and Biobehavioral Reviews, 36(10), 2037–2039.

Lyamin, O.I., Manger, P.R., Ridgway, S.H., Mukhametov, L.M. and Siegel, J.M. (2008). 'Cetacean sleep: an unusual form of mammalian sleep.' Neuroscience and Biobehavioral Reviews, 32(8), 1451–1484.

Rattenborg, N.C., Amlaner, C.J. and Lima, S.L. (2000). 'Behavioral, neurophysiological and evolutionary perspectives on unihemispheric sleep.' Neuroscience and Biobehavioral Reviews, 24(8), 817–842.

Rattenborg, N.C., Lima, S.L. and Amlaner, C.J. (1999). 'Facultative control of avian unihemispheric sleep under the risk of predation.' Behavioural Brain Research, 105(2), 163–172.

Rattenborg, N.C., Voirin, B., Cruz, S.M., Tisdale, R., Dell'Omo, G., Lipp, H.P., Wikelski, M. and Vyssotski, A.L. (2016). 'Evidence that birds sleep in mid-flight.' Nature Communications, 7, 12468.

Shaw, N.D., McHill, A.W., Sato, S., Dobrosielski, D.A., Riera, C.E., Siegel, J.M. and Czeisler, C.A. (2022). 'Acute sleep disruption does not diminish pulsatile growth hormone secretion in pubertal children.' Journal of the Endocrine Society, 6(11), bvac142. PMC9562791.

Siegel, J.M. (2009). 'Sleep viewed as a state of adaptive inactivity.' Nature Reviews Neuroscience, 10(10), 747–753.

Takahashi, Y., Kipnis, D.M. and Daughaday, W.H. (1968). 'Growth hormone secretion during sleep.' Journal of Clinical Investigation, 47(9), 2079–2090.

Xie, L., Kang, H., Xu, Q., Chen, M.J., Liao, Y., Thiyagarajan, M., O'Donnell, J., Christensen, D.J., Nicholson, C., Iliff, J.J., Takano, T., Deane, R. and Nedergaard, M. (2013). 'Sleep drives metabolite clearance from the adult brain.' Science, 342(6156), 373–377.

Young, S. (2026). 'Natural Healing: Understanding Recovery Across Physical, Psychological, and Therapeutic Domains.' YoungFamilyLife Ltd.

Zaffanello, M., Cancarini, P., Pietrobelli, A., Pecoraro, L. and Franchini, M. (2023). 'Complex relationship between growth hormone and sleep in children: insights, discrepancies, and implications.' Frontiers in Endocrinology, 14. PMC10847528.