Biphasic Sleep, the Two-Sleep Pattern and Where It Stops Being Legitimate

Summary

Biphasic sleep — two sleep episodes across 24h, whether a siesta pattern or a Wehr-style segmented night — is a physiologically real and historically documented pattern that is benign or mildly beneficial only when the two episodes SUM to an adequate total; it is not evidence that you can need less sleep (that is polyphasic hacking, a safety wall), and its reassuring "your 3am waking is natural" reframe is a genuine gift to the anxious normal-waker AND a trap if it talks a real insomnia or apnoea sufferer out of treatment.

Why Moderate

This entry is Moderate overall because its load-bearing "is it BETTER / is it the natural default" claims are not established, even though its "does this pattern exist" claims are strong.

Why not the tier above (Strong). The claims that would justify a higher tier — that biphasic sleep is superior, or that it is the universal human default — are contested (Boyce) and directly counter-exampled (Siegel's monophasic hunter-gatherers). The siesta benefit rests on one confounded observational cohort. The reframe-as-treatment has no dedicated RCT. Strong-tier would over-state what the packaging and default claims can bear.

Why not the tier below (Emerging). The physiological reality of segmented sleep (Wehr), the historical record (Ekirch), the total-sleep principle (Mollicone/Van Dongen/Dinges), the normal-waking baseline (textbook physiology), and the polyphasic safety wall (NSF consensus panel) are all well-established. The entry is not speculative; it is a well-grounded synthesis whose headline "better?" question is genuinely open.

Per-sub-area tier split:
• Physiological basis of segmented sleep (Wehr) — Strong.
• The prolactin/meditative-quality detail — Moderate (qualitative, spread across the oeuvre).
• Historical record of two-sleep pattern (Ekirch) — Strong as history.
• "Universal biological default" extrapolation — Emerging/contested.
• Siesta as a documented cultural pattern — Strong (descriptive).
• Siesta coronary-mortality benefit — Moderate (single confounded cohort).
• Total sleep, not packaging, sets need (Mollicone et al.) — Strong.
• Brief calm awakening is normal — Foundational.
• Polyphasic hacking has no benefit / is harmful — Strong (safety wall).
• The reframe reduces sleep-effort anxiety as a treatment — Emerging (mechanistic, no RCT).

Practical takeaway

The framing to hold. Biphasic sleep is a legitimate way to package an adequate total, not a way to need less. Judge any two-episode pattern by one question first: do the episodes sum to enough total sleep? If yes (siesta plus a solid night; a segmented night that still totals ~7-9h), it is benign and may suit shift workers, parents, siesta cultures, and long-night seasons. If the "pattern" is really a plan to sleep two to four hours total, it is not biphasic sleep — it is chronic deprivation, and the answer is no.

Concrete protocols.
• Siesta as a supplement: a scheduled afternoon sleep is fine when night sleep stays adequate; treat it as an addition to, not a replacement for, the night. (For nap duration, timing, and the fatigue-signal reading, see napping_evidence_and_protocol — this entry does not set nap length.)
• Segmented night, deliberately: if long nights or lifestyle push you into two bouts with a calm interval between, that interval does not need to be "fixed." Keep the room dark and low-stimulation, avoid screens and clock-watching, and let the second sleep arrive. The goal is an adequate total across both bouts, not a forced single block.
• The 3am reframe, used carefully: for someone who wakes calmly, briefly, and returns to sleep, the knowledge that this is physiologically normal can lower the anxiety that would otherwise keep them up. Offer it as reassurance for a calm waker — never as a treatment claim.

Who should not adopt this / when it's a signal. Do not treat the reframe as a reason to ignore waking that is distressing, prolonged, or accompanied by unrefreshing sleep, loud snoring, gasping, or daytime impairment — those are signals, not "ancestral rhythm," and route to sleep_maintenance_insomnia and obstructive-sleep-apnoea screening. Do not adopt polyphasic hacking schedules at all. Anyone using a "biphasic" plan to justify a sub-adequate total is not doing biphasic sleep and should be redirected to total-sleep adequacy.

Evidence detail

Why This Entry Exists

Biphasic sleep is one of the most rhetorically abused topics in the sleep space, and the abuse runs in opposite directions at once. On one side, "ancestral sleep" influencers and biohacking sellers borrow the genuine science of segmented and siesta sleep to imply that consolidated eight-hour sleep is unnatural and that you could — or should — carve your night into pieces and reclaim waking hours. On the other side, a rigid sleep-hygiene orthodoxy treats any nocturnal awakening as a defect to be medicated, pathologising a pattern that Thomas Wehr's photoperiod work and Roger Ekirch's historical scholarship both show can be entirely physiological and calm.

The honest position sits in a narrow band between those two distortions, and it is defined by two hard lines. The first is a total-sleep line: legitimate biphasic sleep means two episodes that sum to an adequate 24h total, and it is categorically different from polyphasic "sleep-hacking" schedules (Uberman, Everyman, Dymaxion) that claim to thrive on two to four hours total — those are chronic sleep deprivation dressed as optimisation, with no evidence of adaptation. The second is a clinical line: the segmented-sleep reframe genuinely reduces sleep-effort anxiety, but a calm, brief waking is not the same phenomenon as distressing, prolonged, symptom-laden waking, and the reframe must never become a reason to self-diagnose away real sleep-maintenance insomnia or obstructive sleep apnoea.

This entry exists because a user who reads only the thesis ("biphasic sleep is natural") will draw exactly the wrong conclusion in at least one direction — either that they can cheat their sleep need, or that their genuinely broken sleep is fine. The entry's job is to hold all of it at once.

What bad advice this protects against, in all directions:
• "Humans are naturally biphasic, so consolidated 8-hour sleep is unnatural and you should split your night." → The historical/segmented pattern is real but culturally and photoperiodically situated, not a proven universal biological default; modern hunter-gatherer field studies find largely monophasic sleep.
• "Polyphasic schedules let you adapt to 2-4 hours of sleep and reclaim your day." → No evidence of adaptation exists; the "adapted" feeling is REM-rebound survival physiology while objective performance degrades. This is chronic partial sleep deprivation, full stop.
• "Your 3am waking is just ancestral segmented sleep, nothing to worry about." → True for a calm, brief, non-distressing waking; dangerously false if the waking is prolonged, distressing, unrefreshing, or accompanied by snoring/gasping — those route to insomnia or apnoea assessment, not self-soothing.
• "Every night waking is a sleep disorder that needs a sleep aid." → Brief arousals at sleep-cycle boundaries are normal textbook physiology and increase benignly with age; pathologising all of them feeds orthosomnia and unnecessary hypnotic use.
• "A siesta protects your heart, so nap for health." → The siesta-mortality association is a single confounded observational cohort; the nap as a fatigue signal and the long-nap harm epidemiology are a different literature owned by a neighbouring entry.

This entry OWNS biphasic sleep as a scheduled sleep architecture — the deliberate two-episode pattern (siesta cultures, Wehr-style segmented nights), the historical/segmented-sleep reframe, the total-sleep dividing line, and the polyphasic safety wall. It DEFERS the nap-as-recovery-tool, the power-nap protocol, and the long-nap harm epidemiology to napping_evidence_and_protocol; it DEFERS the treatment of distressing middle-of-the-night waking to sleep_maintenance_insomnia; and it DEFERS the mechanics of sleep cycles and arousals to sleep_architecture_and_stages.

Evidence

Read the tiers, not the thesis. The historical and physiological "this pattern exists" claims are strong; the "this pattern is the natural default" and "the reframe treats insomnia" claims are only Emerging; the siesta-benefit claim is Moderate and confounded; the polyphasic-harm claim is strong enough to be a safety wall. The thesis you should carry away is not "biphasic is better" — it is "two adequate episodes are fine, fewer total hours are not."

Sub-area 1 — The physiological basis (segmented sleep is real).

1. (Strong Evidence — small, artificial-photoperiod experiment) Under a shortened photoperiod (~14h of darkness per night), healthy subjects spontaneously settle into a segmented, bimodal sleep pattern: two sleep bouts of several hours separated by a 1-3h interval of quiet wakefulness, with expanded nocturnal melatonin secretion indicating genuine circadian control. Wehr TA, "In short photoperiods, human sleep is biphasic," Journal of Sleep Research 1992;1(2):103-107. A small group of healthy men; total sleep remained adequate (~8h split in two). (Non-commercial NIMH research; cui bono runs toward over-claiming advocates who cite it as proof of a "natural default," and against a rigid single-block orthodoxy that would dismiss it as an artifact.)

2. (Moderate — qualitative/endocrine detail, spread across Wehr's oeuvre) During the mid-sleep waking interval, subjects showed elevated prolactin and described the wakefulness as calm and meditative rather than anxious or distressed — unlike the wakefulness of clinical insomnia. Attributed to Wehr's body of work on this paradigm rather than to one line of the 1992 abstract; no specific prolactin figure is cited here because none was verified. (Over-romanticising the interval as mystical clarity benefits "sacred night waking" wellness content; erasing it entirely benefits sleep-hygiene orthodoxy that medicates any waking.)

Sub-area 2 — The historical record (the "two sleeps").

3. (Strong Evidence — as social history, not clinical proof) Pre-industrial Western European households commonly retired a couple of hours after dusk, woke for one to two hours of quiet wakefulness ("the watch," dorveille), then had a "second sleep" until dawn. A. Roger Ekirch, "Sleep We Have Lost: Pre-industrial Slumber in the British Isles," The American Historical Review 106(2):343-386, April 2001; and Ekirch, At Day's Close: Night in Times Past, W. W. Norton, 2005. The scale is large — more than 500 references drawn from diaries, court depositions, medical texts, prayer manuals, and imaginative literature. (Strong here means strong as history — Ekirch and publishers benefit from a compelling "lost natural rhythm" narrative, and the article's peer-reviewed status is often invoked to lend clinical authority the paper never claims.)

4. (Emerging — the "universal biological default" extrapolation is contested) That segmented sleep is the proven, universal human default and consolidated sleep is "unnatural" is overstated. Niall Boyce, "Have we lost sleep? A reconsideration of segmented sleep in early modern England," Medical History, 2023, argues the sources are fragmentary and ambiguous and that broken sleep often reflected specific circumstances rather than a normalised universal routine (Ekirch replied in Medical History, 2024). Jerome Siegel's field studies of modern pre-industrial groups (Hadza, San, Tsimane; e.g. Yetish et al., Current Biology 2015) found largely CONTINUOUS ~5.7-7.1h sleep — monophasic — a direct counter-example. (Over-claiming the default benefits wellness/reframe sellers; under-claiming benefits a tidy 8-hour-block orthodoxy and its product lines.)

Sub-area 3 — Siesta (culturally scheduled biphasic).

5. (Strong Evidence — descriptive/ethnographic) The siesta pattern (core night sleep plus a scheduled afternoon sleep) is a well-documented normal cultural pattern across the Mediterranean, Latin America, and parts of Asia — two episodes across 24h, legitimate only when they sum to an adequate total. This is a descriptive claim, not an efficacy claim. (Siesta-romanticism over-sells it as inherently superior; hustle culture dismisses it as laziness — neither distortion is in the anthropological record.)

6. (Moderate — single confounded observational cohort) In the Greek EPIC cohort, siesta was associated with lower coronary mortality. Naska A, Oikonomou E, Trichopoulou A, Psaltopoulou T, Trichopoulos D, "Siesta in Healthy Adults and Coronary Mortality in the General Population," Archives of Internal Medicine 2007;167(3):296-301. 23,681 adults free of CHD/stroke/cancer at baseline, mean 6.3y follow-up, 133 coronary deaths: ANY siesta mortality ratio 0.66 (95% CI 0.45-0.97, ~34% lower); SYSTEMATIC siesta (≥3x/wk, ≥30 min) 0.63 (95% CI 0.42-0.93, ~37% lower); OCCASIONAL siesta 0.88 (95% CI 0.48-1.60, non-significant, CI crosses 1). Adjusted for age, sex, smoking, education, BMI, waist-hip ratio, physical activity, and Mediterranean-diet score; strongest in working men. Association, not causation — reverse causation and a healthy-worker effect are live. The two figures (~34% for any, 37% for systematic) belong to different sub-groups and must not be conflated into one number. (Over-claim: "blue zones"/wellness marketing cites 37% as if a nap causally protects the heart. Under-claim: productivity culture dismisses siesta entirely. The truth is a real but confounded inverse association from one cohort.)

Sub-area 4 — The safety wall (polyphasic hacking).

7. (Strong Evidence — consensus review, against the sellers) Polyphasic sleep-hacking schedules (Uberman ~6x20-min naps ≈2h total; Everyman; Dymaxion) that claim you can thrive on 2-4h total sleep have no scientific support. Weaver MD, Sletten TL, Foster RG, Gozal D, Klerman EB, Rajaratnam SMW, et al., "Adverse impact of polyphasic sleep patterns in humans: Report of the National Sleep Foundation sleep timing and variability consensus panel," Sleep Health 2021;7(3):293-302 (PMID 33795195): 40,672 records screened, 22 retained; no evidence of benefit; consensus that polyphasic schedules and their inherent sleep deficiency are associated with adverse physical-health, mental-health, and performance outcomes; explicit recommendation against schedules that reduce or over-fragment 24h sleep. The claimed "adaptation" is REM-rebound survival physiology, not adaptation. Independent expert critique: Piotr Woźniak, "Polyphasic Sleep: Facts and Myths," SuperMemo (self-published, non-peer-reviewed). (Cui bono is one-directional here: productivity-hack and "gain 20+ waking hours" sellers profit from over-claiming benefit; essentially no one profits from exaggerating polyphasic harm. That asymmetry is why it is a wall, not a debate.)

Sub-area 5 — The organising principle and the normal-waking baseline.

8. (Strong Evidence — lab response-surface mapping) Splitting sleep does NOT reduce total sleep need. Mollicone DJ, Van Dongen HPA, Rogers NL, Dinges DF, "Response surface mapping of neurobehavioral performance: Testing the feasibility of split sleep schedules for space operations," Acta Astronautica 2008;63(7-10):833-840, found neurobehavioral impairment accumulates near-linearly with TOTAL 24h time-in-bed, whether consolidated in one nocturnal block or split into an anchor sleep plus a nap. An adequate total split into night + nap preserves performance; a reduced total degrades it either way. (Refutes "polyphasic = need less sleep." Over-claim benefits sleep-hacking gurus; under-claim — "only one block ever works" — disadvantages shift workers, parents, and siesta cultures for whom a well-summed split is legitimately fine.)

9. (Foundational — established textbook physiology) A brief, calm nocturnal awakening is normal, not pathological. Sleep proceeds in ~90-min NREM-REM cycles; lighter stages recur and brief awakenings cluster near REM transitions and cycle boundaries, increasing benignly with age. The distinguishing feature of insomnia disorder is distress, difficulty returning to sleep, prolonged wakefulness, and daytime impairment — not the awakening itself. Standard references, e.g. Colten HR, Altevogt BM (eds), Institute of Medicine, "Sleep Physiology," in Sleep Disorders and Sleep Deprivation: An Unmet Public Health Problem (National Academies Press, 2006). (Over-normalising benefits anti-medicalisation content but delays diagnosis of real apnoea/insomnia; pathologising every waking benefits hypnotic-drug and sleep-tracker markets that sell "perfect uninterrupted sleep" and profit from orthosomnia.)

Mechanism

Photoperiodic control of sleep length and structure. Sleep timing and consolidation are shaped by the duration of the daily dark period via the circadian melatonin signal. When the night is artificially lengthened (Wehr's 14h-dark condition), melatonin secretion expands and sleep spreads out, and rather than remaining one long block it divides into two bouts with a quiet interval between — the same architecture the historical record describes. This is why segmented sleep is best understood as a latent circadian capacity that surfaces under long nights, not a defect and not a proven modern optimum.

Homeostatic sleep need is set by total, not packaging. The pressure to sleep (process S) accumulates with wakefulness and dissipates with sleep roughly as a function of total sleep obtained. Because impairment tracks total 24h sleep near-linearly regardless of how it is split, distributing an adequate total across two episodes leaves the homeostatic account balanced, while cutting the total — no matter how cleverly fragmented — leaves a debt that the body pays in degraded performance and REM/slow-wave rebound. Episode count is a red herring; total physiological need (~7-9h/24h for most adults) is the invariant.

The quiet-wakefulness interval is a low-arousal state, not clinical wakefulness. In the segmented pattern the between-sleeps interval is accompanied by elevated prolactin and is experienced as calm and meditative. The clinically important contrast is arousal level: insomnia is maintained by a cognitive-arousal loop — worry and effort about being awake raise physiological arousal and block the return to sleep. A calm interval does not trigger that loop; a catastrophised one does. This is the mechanism by which a reframe ("this waking is normal") can reduce sleep-effort anxiety — it lowers the cognitive arousal that would otherwise convert a benign surfacing into a maintained awakening.

Normal arousals at cycle boundaries. Human sleep surfaces toward lighter stages near REM transitions and cycle boundaries, producing brief awakenings that most people do not remember. This baseline is why a single calm waking is architecturally unremarkable, and why frequency, duration, return latency, distress, and daytime consequence — not the mere fact of waking — are the features that separate benign surfacing from disorder.

Risks And Contraindications

• The masking risk (both ways). The segmented-sleep reframe can reassure a benign waker AND conceal a treatable disorder. If waking is distressing, prolonged, frequent, or symptom-laden (snoring, gasping, unrefreshing sleep, daytime sleepiness), do not self-soothe it as "just segmented sleep" — screen for insomnia and apnoea.
• Do not over-claim the default. Segmented/biphasic sleep is a real and historically documented pattern; it is NOT an established universal biological default, and consolidated sleep is not "unnatural." Present it as an option and a normalisation, not a prescription.
• Do not over-claim siesta benefit. The coronary-mortality association is a single confounded observational cohort. Do not tell anyone a nap will protect their heart.
• The total-sleep guardrail. Any two-episode plan that reduces total sleep below individual need carries the full risks of chronic partial sleep deprivation. Splitting does not buy free sleep.
• Polyphasic hacking is contraindicated. Uberman/Everyman/Dymaxion-style schedules impose chronic deprivation with cognitive, mood, metabolic, and safety costs (including drowsy-driving risk). There is no adaptation to be reached.
• Individual variation. Sleep need varies; "adequate total" is a per-person judgement, not a universal number, and extreme fragmentation can degrade slow-wave and REM sleep even at borderline-adequate totals.

Controversy

Nature of the dispute. The core disagreement is whether biphasic/segmented sleep is the natural human default (from which modern consolidated sleep is a deviation) or a historically and photoperiodically contingent pattern that is real but not privileged. A second dispute layers on top: whether the "your waking is natural" reframe is a helpful normalisation or a risky dismissal of real pathology.

Position A — biphasic sleep is the natural default (where it has its best evidence, and where it overreaches). Its best evidence is genuine: Wehr's photoperiod experiment shows the segmentation is circadian and physiological, and Ekirch's large historical record shows two-sleep patterns were common in pre-industrial Europe. Where it overreaches is the jump from "documented and physiologically possible" to "universal default, and consolidated sleep is unnatural." Wehr used an artificial 14h-dark condition unlike modern life, and Siegel's hunter-gatherer field data show largely monophasic sleep — a direct counter-example the "default" claim cannot absorb.

Position B — consolidated sleep is the norm and night waking is a problem (where it has its best evidence, and where it overreaches). Its best evidence is clinical: sleep-maintenance insomnia and obstructive sleep apnoea are real disorders, and consolidated sleep is what most modern adults under normal photoperiods actually get. Where it overreaches is treating ALL night waking as pathological — brief arousals at cycle boundaries are normal textbook physiology, they increase benignly with age, and pathologising them drives orthosomnia, unnecessary hypnotics, and anxiety that itself worsens sleep.

The funding/bias dimension — cui bono, both ways. Over-claiming the biphasic default benefits "ancestral sleep" influencers, reframe sellers, sleep coaches, and mattress/product marketers who profit from framing modern sleep as broken and their fix as the return to nature — and, at the extreme, polyphasic-hacking course and app sellers who borrow siesta's legitimacy to sell sleep deprivation as productivity. Under-claiming — insisting only one uninterrupted block is healthy — benefits hypnotic-drug makers and sleep-tracker vendors who profit from anxiety about normal arousals, and a tidy orthodoxy that disadvantages shift workers, parents, and siesta cultures. The one place the incentive is NOT symmetric is polyphasic hacking: sellers profit from exaggerating its benefit, and essentially no one profits from exaggerating its harm — which is exactly why the mainstream science lines up against the sellers there.

Realised Position: Biphasic sleep is a real, benign pattern when — and only when — its two episodes sum to an adequate total; it is neither a mandated natural ideal nor a disorder. We use the segmented-sleep reframe as reassurance for a calm, brief waker, and we refuse to let it mask distressing or symptom-laden waking, which we route to clinical assessment. We treat polyphasic hacking as a safety wall, not a lifestyle option. The dividing line we hold is total sleep obtained, not episode count. The honest synthesis sells nothing: it does not sell you a nap for your heart, a reframe as a cure, or a schedule that reclaims your night.

Cross-Pillar Connections

• napping_evidence_and_protocol — Owns the nap as a recovery tool and fatigue signal, the power-nap protocol, and the long-nap harm epidemiology (nap-duration mortality/CVD associations). This entry defers all nap-length and nap-harm material there and does not restate those numbers; here the nap matters only as one episode of a scheduled two-episode architecture.
• sleep_maintenance_insomnia — Owns the assessment and treatment of distressing, prolonged, or recurrent middle-of-the-night waking. This entry defers every distressed or symptom-laden waking there; it keeps only the calm, brief, benign waking and the reframe that reassures it.
• sleep_architecture_and_stages — Owns the NREM-REM cycle structure, stage definitions, and the mechanics of arousals. This entry defers the architecture detail there and uses only the boundary fact that brief awakenings at cycle transitions are normal.
• circadian_rhythm_optimization — Owns light exposure, melatonin timing, and circadian alignment as levers. This entry defers the "how to shift/align your clock" mechanics and uses only the photoperiodic control that makes segmented sleep a latent circadian capacity.
• sleep_effort_and_clock_watching — Owns the cognitive-arousal / sleep-effort loop and clock-watching. This entry defers the anxiety-loop mechanism there and uses it only to explain why a calm reframe can help and a catastrophised waking cannot.
• sleep_debt_payback — Owns accumulated sleep debt and its repayment dynamics. This entry defers debt mechanics there and uses only the corollary that a split which reduces total sleep incurs debt like any other shortfall.
• chronotypes_morningness_eveningness — Owns individual timing preference and its biology. This entry defers chronotype variation there and notes only that "adequate total" and workable episode timing are individual judgements, not universal numbers.

What would change our mind

Industry bias note

Sources (11)

• Wehr TA. "In short photoperiods, human sleep is biphasic." Journal of Sleep Research. 1992;1(2):103-107. doi:10.1111/j.1365-2869.1992.tb00019.x.↗ (Non-commercial NIMH research; no product conflict.) Finding: a small group of healthy men on a ~14h-dark photoperiod developed two sleep bouts separated by a 1-3h calm-wakefulness interval, with expanded nocturnal melatonin — a physiological, circadian basis for segmented sleep at an adequate total.
• Wehr TA — associated endocrine reporting on the same paradigm. (Same non-commercial source; qualitative/self-report detail, no product conflict.) Finding: elevated prolactin during the mid-sleep interval and a calm/meditative rather than distressed quality of the wakefulness; attributed generically to Wehr's body of work, with no specific prolactin figure asserted.↗
• Ekirch AR. "Sleep We Have Lost: Pre-industrial Slumber in the British Isles." The American Historical Review. 2001;106(2):343-386. (Academic history; author/publisher benefit from a compelling "lost rhythm" narrative.) Finding: pre-industrial Western Europeans commonly slept in two intervals ("first sleep," "second sleep") bridged by up to an hour or more of quiet wakefulness.↗
• Ekirch AR. At Day's Close: Night in Times Past. W. W. Norton; 2005. (Trade book; same narrative interest.) Finding: book-length documentation of segmented sleep drawing on 500+ historical references — strong as social history, not as clinical proof of a biological default.↗
• Boyce N. "Have we lost sleep? A reconsideration of segmented sleep in early modern England." Medical History. 2023. (Academic critique; corrective interest.) Finding: the historical sources are fragmentary and ambiguous, and segmented sleep was often contextual rather than a normalised universal routine (Ekirch replied in Medical History, 2024).↗
• Yetish G, et al. (Siegel lab). "Natural sleep and its seasonal variations in three pre-industrial societies." Current Biology. 2015. (Academic field study; no product conflict.) Finding: Hadza, San, and Tsimane groups sleep ~5.7-7.1 largely continuous (monophasic) hours — a direct counter-example to the "biphasic is the innate default" claim.↗
• Naska A, Oikonomou E, Trichopoulou A, Psaltopoulou T, Trichopoulos D. "Siesta in Healthy Adults and Coronary Mortality in the General Population." Archives of Internal Medicine. 2007;167(3):296-301. (Academic cohort; no commercial siesta interest, but frequently cited by wellness marketing.) Finding: in 23,681 Greek adults (mean 6.3y, 133 coronary deaths), any siesta MR 0.66 (95% CI 0.45-0.97, ~34% lower) and systematic siesta MR 0.63 (95% CI 0.42-0.93, ~37% lower); occasional siesta non-significant — an association, heavily confounded by reverse causation and a healthy-worker effect.↗
• Mollicone DJ, Van Dongen HPA, Rogers NL, Dinges DF. "Response surface mapping of neurobehavioral performance: Testing the feasibility of split sleep schedules for space operations." Acta Astronautica. 2008;63(7-10):833-840. (Space-operations research funding; no consumer-product conflict.) Finding: neurobehavioral impairment accumulates near-linearly with TOTAL 24h time-in-bed regardless of split vs consolidated — splitting an adequate total preserves performance; reducing the total does not.↗
• Weaver MD, Sletten TL, Foster RG, Gozal D, Klerman EB, Rajaratnam SMW, et al. "Adverse impact of polyphasic sleep patterns in humans: Report of the National Sleep Foundation sleep timing and variability consensus panel." Sleep Health. 2021;7(3):293-302. pubmed.ncbi.nlm.nih.gov/33795195↗/" target="_blank" rel="noopener">PMID 33795195↗. (Expert consensus panel; the incentive here runs against the sellers, not with them.) Finding: 40,672 records screened, 22 retained; no evidence of benefit from polyphasic schedules and consensus for adverse health/mental-health/performance outcomes; explicit recommendation against schedules that reduce or over-fragment 24h sleep.
• Woźniak P. "Polyphasic Sleep: Facts and Myths." SuperMemo (self-published, non-peer-reviewed). (Independent expert commentary; author sells a spaced-repetition product, not a sleep schedule.) Finding: no correspondent sustained a true Uberman schedule long-term; humans are biphasic and forced polyphasic schedules cost health, time, and cognition.↗
• Colten HR, Altevogt BM, eds. "Sleep Physiology," in Sleep Disorders and Sleep Deprivation: An Unmet Public Health Problem. Institute of Medicine; National Academies Press; 2006. (Public-health reference; no product conflict.) Finding: brief arousals and awakenings are a normal architectural feature of sleep, clustering near cycle boundaries and increasing benignly with age; pathology is defined by distress, prolonged wakefulness, and daytime impairment, not by the awakening itself.↗

Related entries