Vitality Tracker
Request an invitation
Science & Knowledge Hub · Article · 19 min read

Optimal sleep: what the evidence supports, and what it does not

Ask what optimal sleep is and the field answers with more confidence than it has earned. Most of the numbers consumer sleep technology puts in front of you, deep-sleep percentages, cycle-timed alarms, readiness scores, have no demonstrated relationship to any health outcome, and the ones that do, duration, regularity and timing, are the ones the apps foreground least. This article sets out how sleep quality is actually defined, which measures predict outcomes, what a device on your wrist or finger can and cannot know, which popular claims fail, and the definition of optimal sleep that survives, which is the one Vitality Tracker builds on.

Csanad Novak8 October 2026Version 1.0
On this page

How the field defines sleep quality

There is no laboratory assay for sleep quality. What exists is a set of expert frameworks, and it matters what kind of knowledge each one is. The thresholds in nearly every sleep app trace to a 2017 National Sleep Foundation panel: eighteen experts screened 3,928 papers down to 277 and then voted, using a formal appropriateness method, on what counts as good sleep. Falling asleep in 15 minutes or less, no more than one awakening of five minutes or more (two, over 65), 20 minutes or less awake after sleep onset, and sleep efficiency of 85 percent or better are the results. They are structured expert opinion, not cut-points derived from health outcomes, and the panel could reach no consensus on arousals at any age. The ubiquitous 85 percent efficiency target comes from that vote and from nothing else.

The more important conceptual contribution is Buysse's 2014 sleep-health framework, which broke the field's fixation on duration: sleep health is a pattern of sleep and wakefulness, adapted to a person's demands, that promotes well-being, and it has several dimensions, satisfaction, timing, duration, efficiency and daytime alertness. Regularity was added later. Buysse considered it in 2014 and set it aside as less clearly tied to outcomes; the following decade reversed that judgment. The framework holds up better than the questionnaire built on it, which has unstable psychometrics across cultures and two incompatible scorings in circulation.

On duration, the American Academy of Sleep Medicine and the Sleep Research Society recommend seven hours or more for adults aged 18 to 60. They chose a floor rather than a range deliberately: the panel could find no plausible mechanism by which long sleep would cause harm, so it set no ceiling. That is a more careful position than the popular seven-to-nine, and the genetic evidence below bears it out.

The best normative reference for sleep architecture is a 2019 meta-analysis of 169 laboratory studies and 5,273 healthy adults. Across all adults, total sleep time averaged 395 minutes, efficiency 85.7 percent, 48 minutes awake after sleep onset, 20.4 percent deep sleep and 19.0 percent REM. One widely believed claim fails against it: deep sleep and REM percentages do not change significantly with age. What declines with age is continuity, not architecture: about ten minutes of sleep and two points of efficiency per decade, with ten more minutes awake. The idea that deep sleep collapses with age comes from older data that did not screen out undiagnosed sleep disorders, and a lower deep-sleep share in a 60-year-old is not explained by age alone.

What predicts outcomes

The organizing discipline is to prefer evidence that survives a causal test. The list of sleep findings that do is short. Short sleep raises the risk of myocardial infarction by about a fifth, a result supported by Mendelian randomization in two cohorts and the strongest in sleep epidemiology. A morning chronotype lowers breast cancer risk. Frequent daytime napping and insomnia symptoms raise cardiovascular risk, though the napping instruments deserve scrutiny and the insomnia result concerns the subjective complaint. Cognitive behavioral therapy for insomnia halved incident depression in older adults in one trial and was null in a larger one. Extending sleep by 1.2 hours cut energy intake by about 270 kcal a day over two weeks. And the one randomized test of a sleep intervention against hard cardiovascular endpoints, CPAP for sleep apnea in three trials totaling 4,225 patients, was null, even though it drove the apnea index from 29 events an hour to under four.

Two corrections to the way these findings are usually told. First, no randomized trial of a behavioral sleep intervention, extending duration, shifting timing or improving regularity, has shown a reduction in cardiovascular events or mortality. Every behavioral recommendation on this page, including ours, rests on observational and genetic evidence. Second, the null CPAP trials do not show that sleep apnea is harmless. A null treatment trial admits several explanations: adherence averaged 3.3 hours a night, treatment began late in the disease, the populations were selected, non-sleepy patients with established heart disease. What the trials do show is that CPAP cannot be prescribed to such patients for the purpose of preventing cardiovascular events, and that it reliably improves snoring, sleepiness, mood and quality of life, which is reason enough to find the people who have it.

The duration curve is J-shaped, and its two arms are not equally credible. Across 27 cohorts and 1.4 million people, short sleep carried a relative risk of 1.12 for mortality and long sleep 1.30; the largest meta-analysis of all, 3.3 million people, found no short-sleep mortality signal. The long arm is two to three times larger and steepens monotonically, which is the signature of reverse causation: cancer, heart failure, lung disease, depression and frailty all cause long sleep. When sleep was measured objectively in UK Biobank, the longest-sleeping fifth had lower mortality, not higher. The seven-to-eight-hour nadir is a population figure, not a personal prescription, and no trial has randomized sleep duration and counted deaths.

Regularity is the headline finding and deserves to be read carefully. In 60,977 UK Biobank adults with more than ten million hours of accelerometry and 1,859 deaths over 6.3 years, the Sleep Regularity Index predicted all-cause mortality better than duration did. The most regular fifth had 30 percent lower mortality than the least regular after full adjustment (the widely quoted 48 percent is the minimally adjusted figure), and adding duration to the regularity model did not improve it. In sensitivity analyses the association held after adjusting, one at a time, for sleep timing, prevalent cancer, diabetes, vascular disease, body mass, depressed mood and mental-health visits. Three caveats hold. The claim that regularity beats duration rests on a model comparison and a non-significant test, an argument partly from a null. The top duration fifth in objective data slept only 7.56 hours or more, so duration could not display its U-curve. And the index penalizes wake and napping by construction, so part of what it captures may be fragmentation from illness. The defensible statement is that regularity was the strongest predictor among the objectively measured sleep variables compared in the largest cohort to date, that behaviorally the most regular fifth sleep and wake within about an hour's window, and that it is the sleep behavior most within a person's control.

Sleep architecture carries a signal, but a confounded one. Low REM has the most credible mortality association of any stage: each 5 percent reduction carried a hazard ratio of 1.13 in one cohort and 1.17 in another. But antidepressants, alcohol, opioids, apnea and age all suppress REM, and that confounding is first-order. Falling slow-wave sleep predicted incident dementia in Framingham, on 346 people and 52 events, and the lowest quartile of slow-wave sleep predicted incident hypertension in older men. None of it supports a percentage target, and the one study that measured slow-wave activity against cognitive decline found decline at both low and high levels, an inverted U that defeats any more-is-better framing.

What can be known at home

Almost every consumer tracker reports sensitivity for sleep of 92 to 96 percent, and the figure is close to meaningless: a device that scored every minute as asleep would be 88 percent accurate in a person with 88 percent efficiency. The discriminating statistic is specificity for wake, and in a 2025 laboratory comparison of six devices it ran from 29 to 52 percent, worst in exactly the people who most want the data: older adults, fragmented sleepers, people with insomnia. All but one device overestimated total sleep, every device overestimated efficiency, and every device underestimated wake after sleep onset, one by 48 minutes.

Sleep staging is the least defensible number on the screen. Stage classification runs at roughly 50 to 65 percent epoch accuracy across the major brands, with per-stage errors of 30 to 60 minutes and wide limits of agreement. The biology compounds the instrument: in 94 healthy men recorded over 470 home nights with laboratory-grade EEG, the deep-sleep share varied from night to night by about a fifth of its mean, and sleep latencies and wake time by half, so a single night's figure carries limited information even when measured perfectly. Manufacturer-adjacent validations report agreement in the nineties; an independent clinical replication of the same device reported 53 percent, with the authors concluding that group-level agreement masks individual-level inaccuracy. Both papers cannot be describing the same device to the same person.

Wearables are good at what happens to matter. Timing, duration and regularity aggregated over weeks match or beat research actigraphy. Overnight resting heart rate agrees with electrocardiography to within a beat, and night-averaged heart-rate variability to within a couple of milliseconds; short-window values are noisier, especially in older adults. Overnight respiratory rate is accurate to about one breath a minute and varies within a person by only 4 to 6 percent across two weeks, which makes a sustained rise against your own baseline meaningful even though the absolute number is not.

Where they fail hardest is pulse oximetry, the weakest sensor and the most aggressively marketed. In a clinic comparison of consumer smartwatches, mean absolute errors ran from 2 to 6 percent, and in a small controlled-hypoxia study the best-known watch's error exceeded the regulatory tolerance precisely in the low range where it would matter; skin-pigmentation bias is largely untested in wrist devices. Proprietary sleep and readiness scores are worse, because the evidence is not weak but absent: no proprietary score has a published validation against any external criterion, their weights are undisclosed, and they change silently between firmware versions. The American Academy of Sleep Medicine's position stands: consumer sleep technology cannot be used to diagnose or treat sleep disorders.

Subjective instruments are not redundant with any of this. The Pittsburgh Sleep Quality Index correlates poorly with actigraphy and better with depression scales, which is usually read as a failure of the questionnaire. It is better read as evidence that subjective and objective sleep are different constructs, and that both matter: a person who sleeps eight efficient hours and feels terrible has a real problem no accelerometer will see. In 123 people with insomnia, self-reported sleep ran about an hour short of the laboratory night on average and four in five underestimated it, while good sleepers more often overestimated theirs; some discrepancy is the norm in both groups.

Claims that do not survive scrutiny

ClaimVerdictWhy
You need 20 to 25 percent deep sleep and 20 to 25 percent REMUnsupportedPopulation means reified into targets. No trial has assigned a stage percentage and measured an outcome; the cognition data show an inverted U; a single night's figure is close to noise.
Sleep cycles are 90 minutes, so wake at the end of oneWeak; stage-timed alarms unsupportedAcross 6,064 recorded cycles the median was 96 minutes, not normally distributed, with the first cycle shorter, two to six cycles a night, and more variation within a person than between people. No trial shows smart alarms improve daytime function, and they buy a small reduction in grogginess by cutting total sleep, the one variable with real evidence.
Deep sleep comes early, REM comes lateWell supportedThe two-process model explains it: homeostatic pressure front-loads deep sleep, the circadian clock gates REM toward early morning. Cutting sleep short therefore truncates REM disproportionately.
You can catch up at the weekendMixed; the folk claim failsIn a metabolic-ward trial, weekend recovery sleep repaid about an hour of debt, delayed circadian phase and left insulin sensitivity worse than restriction alone; a 43,880-person cohort found that, under 65, weekend catch-up sleepers had mortality no different from the reference. Catch-up partly repays duration and fails to normalize physiology.
Sleep before midnight is worth doubleUnsupportedSlow-wave activity is front-loaded into the early hours of your sleep episode, indexed to sleep onset and circadian phase, not the clock face. The nearest real finding, lowest cardiovascular incidence with sleep onset between 10 and 11 pm, is a seven-day observational snapshot with a U-shaped curve.
Long sleep is harmfulUnsupported as causalThe largest observational effect in the field and the weakest causal standing. Genetic evidence corroborates short sleep; long sleep could not be tested genetically, and illness causes long sleep.
Keep the bedroom at 18.3 °CUnsupportedNo trial has titrated bedroom temperature to an optimum; 18.3 is a unit conversion of a round 65 °F. The best objective data in older adults put peak efficiency at 20 to 25 °C, and the mechanism points the other way: sleep onset is gated by heat loss from the skin, so a warm bath an hour or two before bed has better support than a cold room.
Blue-light glasses improve sleepUnsupportedThe 2023 Cochrane review of 17 trials found none at low risk of bias and very-low-certainty evidence for sleep; an earlier meta-analysis found the objective sleep-efficiency effect crossing zero while self-reported effects were large, the signature of an expectancy effect.
Sleep washes amyloid from the brainWeak to unsupportedThe 2013 mouse finding is directly contradicted by 2024 work showing brain clearance reduced during sleep and anesthesia; the dispute is unresolved at the level of the primary animal data, and the human link to dementia is observational and bidirectionally confounded.
Your sleep score tells you how you sleptUnsupportedNo proprietary score has any published validation, and there is a documented harm, orthosomnia: people presenting to sleep clinics whose complaint is the tracker's verdict, some made worse by chasing it.

The pattern across the table is the useful finding. The claims that survive are about mechanism: the two-process model, circadian phase, the thermoregulation of sleep onset. Nearly every claim turned into a consumer target has no trial behind it or a failed one, and the variable with the best evidence, total sleep, is the one stage-optimization and smart alarms trade away.

A defensible definition

Optimal sleep for a healthy adult has six traits, re-anchored from Buysse's framework to the outcome evidence. Sufficient: seven hours or more of actual sleep most nights, without anxiety about the upper bound. Regular: a Sleep Regularity Index of 85 or more, which in practice means falling asleep and waking within about an hour's window every day, weekends included. Well-timed: aligned with your own chronotype rather than forced against it, with the clock-time evidence held loosely. Continuous: efficiency of 85 percent or more, onset within 15 to 20 minutes, few sustained awakenings, on the consensus indicators, which are expert opinion but describe what people experience as bad sleep and what degrades with age and illness. Unobstructed: no sleep-disordered breathing, the one place where a clinical pathway with a treatment exists. Restorative: subjectively satisfying, with sustained daytime alertness. Four of the six are reliably measurable at home. Sleep-stage percentages are not on the list, and neither are composite scores: of seven sleep domains tested against mortality in one older cohort, only rhythmicity and continuity were robust, and no study has shown a composite to discriminate better than its best single component.

How Vitality Tracker does it

The platform's sleep model follows the evidence rather than the market. Regularity leads: the Sleep Health Index is a rolling 28-night figure in which regularity carries the largest weight, duration the second, and continuity, circadian alignment and your own ratings of satisfaction and alertness the rest. Sleep stages are shown as a low-confidence estimate and carry no weight, and the app says why. Every sleep input carries a record of the device that produced it and whether that device has a published calibration; where it does not, continuity is read as a trend against your own baseline rather than against laboratory norms. Overnight heart rate, heart-rate variability and respiratory rate are reported as deviations from your own rolling baseline, because no population reference range exists for them. A score computed from incomplete data shows its coverage and names what is missing. The validated questionnaires clinicians use, STOP-Bang, the Epworth Sleepiness Scale, the Insomnia Severity Index and the Consensus Sleep Diary, are available in the app for you to complete and take to your doctor. The full scoring method, with every weight, curve and limitation and the external review that changed it, is published in the Methods section (doi:10.5281/zenodo.23223863).

Vitality Tracker describes and tracks your sleep pattern, recorded by your own devices and your own entries, against the published evidence and your own history, for general wellness and for discussion with a health professional. It tracks measures, not disease, and does not diagnose, predict or monitor any disease; it cannot rule out sleep apnea, and a sleep complaint belongs with your doctor.

Key takeaways
  • Sleep quality is defined by expert consensus, not by outcome-derived thresholds; treat the targets as a scaffold.
  • Duration, regularity and timing carry the outcome evidence and are measurable at home; sleep stages and composite scores carry neither.
  • Regularity was the strongest predictor among objectively measured sleep variables in the largest cohort to date, and it is the behavior most within your control.
  • The long-sleep hazard is most likely reverse causation; the short-sleep hazard is causal.
  • Wearables are reliable for timing, duration, regularity, overnight heart rate and respiratory rate; they are unreliable for staging, wake detection and oxygen saturation, and no proprietary score has been validated.
  • Subjective sleep is a separate dimension from measured sleep, and both matter.

References

  1. Windred DP, Burns AC, Lane JM, et al. Sleep regularity is a stronger predictor of mortality risk than sleep duration: a prospective cohort study. Sleep 2024;47(1):zsad253.
  2. Daghlas I, Dashti HS, Lane J, et al. Sleep duration and myocardial infarction. Journal of the American College of Cardiology 2019;74(10):1304–1314.
  3. Ohayon M, Wickwire EM, Hirshkowitz M, et al. National Sleep Foundation's sleep quality recommendations: first report. Sleep Health 2017;3(1):6–19.
  4. Buysse DJ. Sleep health: can we define it? Does it matter? Sleep 2014;37(1):9–17.
  5. Watson NF, Badr MS, Belenky G, et al. Recommended amount of sleep for a healthy adult: a joint consensus statement of the American Academy of Sleep Medicine and Sleep Research Society. Sleep 2015;38(6):843–844.
  6. Boulos MI, Jairam T, Kendzerska T, Im J, Mekhael A, Murray BJ. Normal polysomnography parameters in healthy adults: a systematic review and meta-analysis. Lancet Respiratory Medicine 2019;7(6):533–543.
  7. McEvoy RD, Antic NA, Heeley E, et al. CPAP for prevention of cardiovascular events in obstructive sleep apnea. New England Journal of Medicine 2016;375(10):919–931.
  8. Sánchez-de-la-Torre M, Sánchez-de-la-Torre A, Bertran S, et al. Effect of obstructive sleep apnoea and its treatment with continuous positive airway pressure on the prevalence of cardiovascular events in patients with acute coronary syndrome (ISAACC study): a randomised controlled trial. Lancet Respiratory Medicine 2020;8(4):359–367.
  9. Irwin MR, Carrillo C, Sadeghi N, Bjurstrom MF, Breen EC, Olmstead R. Prevention of incident and recurrent major depression in older adults with insomnia: a randomized clinical trial. JAMA Psychiatry 2022;79(1):33–41.
  10. Tasali E, Wroblewski K, Kahn E, Kilkus J, Schoeller DA. Effect of sleep extension on objectively assessed energy intake among adults with overweight in real-life settings: a randomized clinical trial. JAMA Internal Medicine 2022;182(4):365–374.
  11. Cappuccio FP, D'Elia L, Strazzullo P, Miller MA. Sleep duration and all-cause mortality: a systematic review and meta-analysis of prospective studies. Sleep 2010;33(5):585–592.
  12. Kwok CS, Kontopantelis E, Kuligowski G, et al. Self-reported sleep duration and quality and cardiovascular disease and mortality: a dose-response meta-analysis. Journal of the American Heart Association 2018;7(15):e008552.
  13. Huang T, Mariani S, Redline S. Sleep irregularity and risk of cardiovascular events: the Multi-Ethnic Study of Atherosclerosis. Journal of the American College of Cardiology 2020;75(9):991–999.
  14. Leary EB, Watson KT, Ancoli-Israel S, et al. Association of rapid eye movement sleep with mortality in middle-aged and older adults. JAMA Neurology 2020;77(10):1241–1251.
  15. Himali JJ, Baril AA, Cavuoto MG, et al. Association between slow-wave sleep loss and incident dementia. JAMA Neurology 2023;80(12):1326–1333.
  16. Fung MM, Peters K, Redline S, et al. Decreased slow wave sleep increases risk of developing hypertension in elderly men. Hypertension 2011;58(4):596–603.
  17. Lucey BP, Wisch J, Boerwinkle AH, et al. Sleep and longitudinal cognitive performance in preclinical and early symptomatic Alzheimer's disease. Brain 2021;144(9):2852–2862.
  18. Schyvens AM, Peters B, Van Oost NC, et al. A performance validation of six commercial wrist-worn wearable sleep-tracking devices for sleep stage scoring compared to polysomnography. SLEEP Advances 2025;6(2):zpaf021.
  19. Chinoy ED, Cuellar JA, Huwa KE, et al. Performance of seven consumer sleep-tracking devices compared with polysomnography. Sleep 2021;44(5):zsaa291.
  20. Miller DJ, Sargent C, Roach GD. A validation of six wearable devices for estimating sleep, heart rate and heart rate variability in healthy adults. Sensors 2022;22(16):6317.
  21. Chouraki A, Tournant J, Arnal P, Pépin JL, Bailly S. Objective multi-night sleep monitoring at home: variability of sleep parameters between nights and implications for the reliability of sleep assessment in clinical trials. Sleep 2023;46(5):zsac319.
  22. Svensson T, Madhawa K, Nt H, Chung UI, Svensson AK. Validity and reliability of the Oura Ring Generation 3 (Gen3) with Oura sleep staging algorithm 2.0 (OSSA 2.0) when compared to multi-night ambulatory polysomnography: a validation study of 96 participants and 421,045 epochs. Sleep Medicine 2024;115:251–263. Comment: Lolli L. Sleep Medicine 2024;117:216; authors' reply, 217–218.
  23. Herberger S, Aurnhammer C, Bauerfeind S, Bothe T, Penzel T, Fietze I. Performance of wearable finger ring trackers for diagnostic sleep measurement in the clinical context. Scientific Reports 2025;15:9461.
  24. Khosla S, Deak MC, Gault D, et al. Consumer sleep technology: an American Academy of Sleep Medicine position statement. Journal of Clinical Sleep Medicine 2018;14(5):877–880.
  25. Grandner MA, Kripke DF, Yoon IY, Youngstedt SD. Criterion validity of the Pittsburgh Sleep Quality Index: investigation in a non-clinical sample. Sleep and Biological Rhythms 2006;4(2):129–139.
  26. Benz F, Riemann D, Domschke K, et al. How many hours do you sleep? A comparison of subjective and objective sleep duration measures in a sample of insomnia patients and good sleepers. Journal of Sleep Research 2023;32(2):e13802.
  27. Wallace ML, Stone K, Smagula SF, et al. Which sleep health characteristics predict all-cause mortality in older men? An application of flexible multivariable approaches. Sleep 2018;41(1):zsx189.
  28. Borbély AA, Daan S, Wirz-Justice A, Deboer T. The two-process model of sleep regulation: a reappraisal. Journal of Sleep Research 2016;25(2):131–143.
  29. Cajochen C, Reichert CF, Münch M, et al. Ultradian sleep cycles: frequency, duration, and associations with individual and environmental factors. A retrospective study. Sleep Health 2024;10(1 Suppl):S52–S62.
  30. Depner CM, Melanson EL, Eckel RH, et al. Ad libitum weekend recovery sleep fails to prevent metabolic dysregulation during a repeating pattern of insufficient sleep and weekend recovery sleep. Current Biology 2019;29(6):957–967.e4.
  31. Åkerstedt T, Ghilotti F, Grotta A, et al. Sleep duration and mortality: does weekend sleep matter? Journal of Sleep Research 2019;28(1):e12712.
  32. Nikbakhtian S, Reed AB, Obika BD, et al. Accelerometer-derived sleep onset timing and cardiovascular disease incidence: a UK Biobank cohort study. European Heart Journal: Digital Health 2021;2(4):658–666.
  33. Baniassadi A, Manor B, Yu W, Travison T, Lipsitz L. Nighttime ambient temperature and sleep in community-dwelling older adults. Science of the Total Environment 2023;899:165623.
  34. Haghayegh S, Khoshnevis S, Smolensky MH, Diller KR, Castriotta RJ. Before-bedtime passive body heating by warm shower or bath to improve sleep: a systematic review and meta-analysis. Sleep Medicine Reviews 2019;46:124–135.
  35. Singh S, Keller PR, Busija L, et al. Blue-light filtering spectacle lenses for visual performance, sleep, and macular health in adults. Cochrane Database of Systematic Reviews 2023;8(8):CD013244.
  36. Shechter A, Quispe KA, Mizhquiri Barbecho JS, Slater C, Falzon L. Interventions to reduce short-wavelength ("blue") light exposure at night and their effects on sleep: a systematic review and meta-analysis. SLEEP Advances 2020;1(1):zpaa002.
  37. Xie L, Kang H, Xu Q, et al. Sleep drives metabolite clearance from the adult brain. Science 2013;342(6156):373–377.
  38. Miao A, Luo T, Hsieh B, et al. Brain clearance is reduced during sleep and anesthesia. Nature Neuroscience 2024;27(6):1046–1050.
  39. Baron KG, Abbott S, Jao N, Manalo N, Mullen R. Orthosomnia: are some patients taking the quantified self too far? Journal of Clinical Sleep Medicine 2017;13(2):351–354.
  40. Kinnunen H, Rantanen A, Kenttä T, Koskimäki H. Feasible assessment of recovery and cardiovascular health: accuracy of nocturnal HR and HRV assessed via ring PPG in comparison to medical grade ECG. Physiological Measurement 2020;41(4):04NT01.
  41. Dial MB, Hollander ME, Vatne EA, Emerson AM, Edwards NA, Hagen JA. Validation of nocturnal resting heart rate and heart rate variability in consumer wearables. Physiological Reports 2025;13(16):e70527.
  42. Liang T, Yilmaz G, Soon CS. Deriving accurate nocturnal heart rate, rMSSD and frequency HRV from the Oura Ring. Sensors 2024;24(23):7475.
  43. Jiang Y, Spies C, Magin J, Bhosai SJ, Snyder L, Dunn J. Investigating the accuracy of blood oxygen saturation measurements in common consumer smartwatches. PLOS Digital Health 2023;2(7):e0000296.
  44. Jiang Y, Spies C, Roghanizad AR, et al. Performance of wearable pulse oximetry during controlled hypoxia induction: instrument validation study. JMIR Formative Research 2026;10:e85253.
  45. Carney CE, Buysse DJ, Ancoli-Israel S, et al. The Consensus Sleep Diary: standardizing prospective sleep self-monitoring. Sleep 2012;35(2):287–302.
  46. Phillips AJK, Clerx WM, O'Brien CS, et al. Irregular sleep/wake patterns are associated with poorer academic performance and delayed circadian and sleep/wake timing. Scientific Reports 2017;7:3216.

Cite and license

Cite as: Novak C. Optimal sleep: what the evidence supports, and what it does not. Vitality Tracker Science, version 1.0, 8 October 2026. vitalitytracker.app/science/optimal-sleep.

Text © Novak Consulting Ltd, licensed under CC BY 4.0: reuse with attribution.

Vitality Tracker describes and tracks physiological measures recorded by your own consumer devices against your own history. It does not diagnose, predict or monitor any disease.

Who we learn from →