Sleep

The short version

I usually sleep 4.5 to 6 hours a night, which works out to three to four 90-minute cycles. This page starts with the evidence I went looking for, then gives the full picture.
The long-sleep association is real. In the big meta-analyses of cohort studies, long sleep went with higher mortality: a relative risk of 1.30 across 1,382,999 people and 1.39 across 5,134,036. These studies are observational.^1,2
The same kind of data puts the lowest risk at about 7 hours, not 6, and short sleep carries risk too. No study I found shows that 6 hours or less is optimal.^3,4,5
Natural short sleepers exist: people with rare inherited gene variants who sleep 4 to 6 hours a night for life and feel well rested. And three preindustrial societies averaged 5.7 to 7.1 hours of actual sleep.^6,7,8
I found no official Navy or SEAL source that says SEALs sleep in 90-minute intervals. The 90-minute figure is general sleep science (a cycle runs 80 to 100 minutes), and the Army's field manual calls for 7 to 9 hours in every 24.^9,10

The protocol

This is what the sources recommend or measured, not a prescription and not a record of my own routine.

Parameter	What the sources say	Source
Lowest-risk duration in cohorts	About 7 hours a day	^3,4
Professional consensus (AASM and SRS)	At least 7 hours a night for adults	¹¹
Sleep cycle	80 to 100 minutes, usually 4 to 6 cycles a night	⁹
Army field manual	7 to 9 hours in every 24 (as reported by Stars and Stripes)	¹⁰
Army napping guidance	Short, infrequent naps when night sleep isn't possible	¹⁰
Waking up	Grogginess is worst out of slow-wave sleep and rarely lasts beyond 30 minutes without major sleep deprivation	¹²

By my arithmetic, three 90-minute cycles is 4.5 hours and four is 6. NHLBI puts a cycle at 80 to 100 minutes, so that count is an approximation, not a fixed clock.⁹

What I went looking for

Moderate Long sleep and mortality. A meta-analysis of 16 prospective studies (27 cohort samples, 1,382,999 people, 112,566 deaths) found a pooled relative risk of death of 1.30 for long sleep and 1.12 for short sleep.¹ A 2018 meta-analysis of 137 cohort studies (5,134,036 people) linked long sleep to higher mortality (relative risk 1.39, 95% CI 1.31 to 1.47) and to diabetes (1.26), cardiovascular disease (1.25), stroke (1.46), and coronary heart disease (1.24), though not to high blood pressure (1.01).² In a meta-analysis of heart outcomes, long sleep carried the higher risk for stroke (1.65 versus 1.15 for short sleep) and for total cardiovascular disease (1.41 versus a nonsignificant 1.03), while short sleep carried the higher risk for coronary heart disease (1.48 versus 1.38).¹³ For type 2 diabetes it was 1.48 for long sleep and 1.28 for short.¹⁴ In 28 articles covering 95,259 older adults, long sleep had an adjusted hazard ratio of 1.24 (95% CI 1.16 to 1.33), while short sleep was only marginal at 1.04; the authors found a high probability of publication bias on the long side, though correcting for it strengthened the result.¹⁵ All of this is observational.

Moderate The case the long-sleep researchers made. The American Cancer Society study of more than 1.1 million adults found the best survival at 7 hours a night. Mortality was significantly higher at 8 hours or more and also at 6 hours or less, and the excess passed 15% above 8.5 hours or below 3.5 or 4.5 hours. The authors called the long-sleep risks slight, wrote that short sleep and insomnia "seem associated with little risk distinct from comorbidities," and said causality is unproven.⁴ In a 2004 review, Youngstedt and Kripke wrote that sleeping longer than 7.5 hours was associated with about 5% of all deaths in those data, even after controlling for 32 possible confounders. They proposed testing whether mild sleep restriction lowers mortality in long sleepers, likened it to dietary restriction, and noted that sleep restriction is perhaps the most effective treatment for primary insomnia and that too much time in bed can bring daytime lethargy and fragmented sleep. They also said epidemiology can't prove that long sleep causes death.¹⁶

Moderate Where the risk bottoms out. A dose-response meta-analysis of prospective cohorts in generally healthy people found U-shaped curves for all-cause mortality and cardiovascular events, with the lowest risk at about 7 hours a day. The relative risk of death was 1.06 for each hour under 7 and 1.13 for each hour over, so the curve climbs faster on the long side, and stroke rose 1.18 per extra hour.³ An umbrella review of 69 meta-analyses found the same shape: each hour less went with a 3% to 11% higher risk of outcomes including death, coronary heart disease, stroke, and type 2 diabetes among short sleepers, and each hour more with a 7% to 17% higher risk of stroke, coronary heart disease, and type 2 diabetes among long sleepers.¹⁷ These studies land on about 7 hours, not 6.

Early Preindustrial sleep. Researchers recorded sleep in three preindustrial societies (the record lists Tanzania, Namibia, and Bolivia). Sleep periods averaged 6.9 to 8.5 hours and actual sleep 5.7 to 7.1 hours, near the low end of industrial societies. People fell asleep on average 3.3 hours after sunset, usually woke before sunrise, near the coldest point of the night, and napped on fewer than 7% of winter days and fewer than 22% of summer days.⁸ The authors read this as the core human sleep pattern. Six hours sits inside it; 4.5 is below the lowest group average. The study measured sleep, not health.

Early Natural short sleepers. The University of California, San Francisco group behind these findings defines natural short sleep as a lifelong tendency to sleep only 4 to 6 hours a night and feel well rested.⁶ The first gene found was DEC2: mice carrying the human mutation slept less⁷, and the group's 2019 paper describes the human carriers as sleeping 6 hours a night for their entire lives "without apparent negative effects."⁶ A second DEC2 variant, in one twin of a non-identical pair, went with shorter sleep, less recovery sleep, and fewer performance lapses during sleep deprivation than the other twin.¹⁸ A very rare mutation in the ADRB1 gene turned up in a family of natural short sleepers, and mice carrying it slept less; the group reports more than 50 such families with a lifelong need of under 6.5 hours.⁶ So natural short sleep is real, and my range sits inside its 4 to 6 hours. But these are rare inherited variants traced through families, cells, and mice. They don't show that anyone can train into short sleep, and I don't know whether I carry one.

Early Cycles and waking. Sleep alternates between non-REM and REM, the cycle starts over every 80 to 100 minutes, there are usually four to six cycles a night, and people may wake briefly between them. Deep, slow-wave sleep comes mostly early in the night and REM mostly later.⁹ So a shorter night cuts mainly the later, REM-heavier cycles. On Patrick's podcast, Matthew Walker estimated that someone who goes to bed at midnight and wakes at 6 instead of 8 loses 25% of total sleep but may lose almost 80% of REM sleep; that's his estimate, not a study result.¹⁹ Waking is groggiest out of slow-wave sleep and least groggy out of stage 1 or 2, with REM in between. That grogginess, called sleep inertia, has lasted anywhere from 1 minute to 4 hours across studies, but it rarely exceeds 30 minutes without major sleep deprivation.¹² So timing a wake-up to the end of a cycle has a real basis: the stage you wake from matters. I found no trial that tested alarms timed to 90-minute multiples, and because cycles vary, a fixed count can miss.

What the military and NASA say

I found no official Navy, Naval Special Warfare, or SEAL source that says SEALs sleep in 90-minute intervals. Here is what I could verify:

The Army's Holistic Health and Fitness field manual, as reported by Stars and Stripes in October 2020, calls for soldiers to get 7 to 9 hours of sleep in every 24 hours. It says "the notion that one can adapt to sleep loss is a myth" and that soldiers can't be trained to perform better on less sleep.¹⁰
The same manual allows short, infrequent naps when night sleep isn't possible, and the longest nap possible as often as time allows when sleep time is hard to predict. During restricted sleep (6 hours or less a night), it says napping combined with appropriate doses of caffeine (about 200 mg, the report adds in brackets) may help sustain cognitive performance and alertness.¹⁰
The one SEAL study I found tested caffeine, not a sleep schedule. In a randomized trial, 68 SEAL trainees took 100, 200, or 300 mg of caffeine or a placebo after 72 hours without sleep and under continuous stress. Sleep loss and stress worsened performance and mood, and 200 and 300 mg improved vigilance, reaction time, and alertness.²⁰
A NASA Ames review found that planned naps improve later performance and alertness in operational settings, including a long-haul flight study in which in-flight naps helped crews during descent and landing, and it cautioned that grogginess after a nap should be weighed before adding naps to work schedules.²¹ That's a nap for tired crews, not a replacement for night sleep.

The full picture

The evidence above is real, and so is the rest.

Moderate (a consensus statement, not a trial). The American Academy of Sleep Medicine and the Sleep Research Society used a modified RAND Appropriateness Method, with a panel vote on the evidence, to reach a consensus recommendation of at least 7 hours a night for adults.^11,22

Moderate Restriction experiments find costs people don't notice. In a randomized lab study of 48 healthy adults aged 21 to 38, 14 days of 4 or 6 hours in bed produced cognitive deficits that built day by day; 6 hours or less a night ended up equal to as much as 2 nights of total sleep deprivation, while sleepiness ratings barely rose after the first days, so people were largely unaware of the decline.²³ In 66 volunteers given 3, 5, 7, or 9 hours in bed for 7 days, speed on a vigilance test fell and then leveled off at a lower level at 5 and 7 hours (lapses rose only at 5), kept falling at 3, and stayed at baseline at 9. Three recovery nights of 8 hours didn't bring the 5- and 7-hour groups back. The authors read the plateau as the brain adapting to chronic restriction, but at a reduced level of performance.²⁴ These are lab studies of 1 to 2 weeks, so they can't say what happens over years.

Moderate Short sleep and disease. In 153 studies of 5,172,710 people, short sleep went with higher mortality (relative risk 1.12), diabetes (1.37), high blood pressure (1.17), cardiovascular disease (1.16), coronary heart disease (1.26), and obesity (1.38), and mortality rose in a straight line as sleep fell below 6 hours.⁵ Short sleep of about 5 to 6 hours or less carried a relative risk of 1.28 for type 2 diabetes¹⁴ and 1.48 for coronary heart disease.¹³ These are observational, like the long-sleep studies.

Early Metabolism under severe restriction. Eleven young men limited to 4 hours in bed for 6 nights had lower glucose tolerance (p < 0.02) and higher evening cortisol (p = 0.0001) than after 6 nights of 12 hours in bed.²⁵ Four hours is below my range, so I read that as what happens at the extreme, not an estimate for 5 or 6 hours.

Mixed Whether long sleep is itself the problem. A critical review of 42 prospective studies from 35 populations found that the literature doesn't support a consistent link between self-reported sleep and mortality, called a robust U-shaped curve premature, and pointed to measurement, response bias, confounding, and reverse causation.²⁶ The authors of the 2017 dose-response analysis note that long sleep travels with fatigue and lethargy, depressive symptoms, low income, unemployment, and less education, all risks in their own right.³ In the Whitehall II cohort of 10,308 British civil servants, sleep that rose from 7 or 8 hours went with more non-cardiovascular deaths (hazard ratio 2.1, 95% CI 1.4 to 3.1), and sleep that fell from 6, 7, or 8 hours went with more cardiovascular deaths (2.4, 95% CI 1.4 to 4.1); adjusting for existing illness and health habits left both largely unchanged.²⁷ The authors of the largest long-sleep meta-analysis wrote that future studies still need to test whether the link is causal and modifiable.² And the cohort that made the 8-hour finding famous says causality is unproven.⁴

What that adds up to: the long-sleep link is an association whose own researchers say cause is still untested. The short side has the same observational problem, but it also has lab experiments that show costs within days.^2,23,24

Where my 4.5 to 6 hours sits

My range is below the lowest-risk point in the cohorts (about 7 hours) and below the consensus floor of 7 hours.^3,11 It sits inside the 4 to 6 hours that defines natural short sleep, which comes from rare inherited variants I may or may not have.⁶ The top of it, 6 hours, is inside the preindustrial range of 5.7 to 7.1 hours of actual sleep, and the bottom, 4.5, is below it.⁸ In the American Cancer Society data, the excess risk passed 15% below 3.5 or 4.5 hours.⁴ No study I found shows that 6 hours or less is optimal. And the lab data say feeling fine isn't a reliable test after restricted sleep.²³

Rhonda Patrick on sleep

Rhonda Patrick, PhD, doesn't argue for short sleep in anything I could read, and she lands on the long side of the range:

In a clip from her appearance on The Mel Robbins Podcast (posted July 16, 2026), she says the first question for someone who thinks they're doing everything right is whether they get 7.5 to 9 hours of actual sleep, not just time in bed, followed by morning light, a steady wake time, not eating within 3 hours of bed, and alcohol.²⁸
In FoundMyFitness episode #45 (February 28, 2019), Matthew Walker says "somewhere between seven to nine" hours is what he recommends and that short sleep predicts all-cause mortality. He also argues that hunter-gatherers aren't a good control for how modern people should sleep.¹⁹
A FoundMyFitness episode from May 30, 2023, says passive heating may improve slow-wave sleep, with example protocols of 20 minutes in a 176°F (80°C) sauna or 20 to 30 minutes in a 104.9°F (40.5°C) hot tub or bath, 1 to 2 hours before bed.²⁹ The heat side is on the sauna page.

I'm not saying she endorses my sleep, this site, or anything I do.

How I use it

I usually get 4.5 to 6 hours of sleep a night, which works out to three to four 90-minute cycles. My bedtime, wake time, naps, and tracking go here next; until then, everything above is what the research says, not my routine.

Cautions

Cohort studies show association on both sides. They don't show that long sleep causes harm, and they don't show that 6 hours is safe.^26,4
Restricted sleep can impair people without their noticing: in the 14-day study, sleepiness ratings didn't keep up with the deficits.²³
The Army's field manual calls adapting to sleep loss a myth, and in the lab, 5 and 7 hours in bed left performance stuck at a lower level that 3 recovery nights didn't fix.^10,24
Short sleep goes with higher rates of diabetes, high blood pressure, and coronary heart disease in large cohorts.^5,13
I am not a doctor, and this is what I read, not medical advice.

References

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.Meta-analysisPubMed 20469800 · DOI
What it found: Meta-analysis of 16 prospective studies (27 cohort samples, 1,382,999 participants, 112,566 deaths). Short sleep carried a pooled relative risk of death of 1.12 and long sleep 1.30, with no sign of publication bias but significant heterogeneity for long sleep.
Limits: Pools self-reported sleep from observational cohorts, so it cannot show that long sleep causes death, and the abstract does not say where in the range risk is lowest.
Jike M, Itani O, Watanabe N, Buysse DJ, Kaneita Y. Long sleep duration and health outcomes: A systematic review, meta-analysis and meta-regression. Sleep Medicine Reviews. 2018;39:25-36.Meta-analysisPubMed 28890167 · DOI
What it found: Meta-analysis of 137 prospective cohort studies with 5,134,036 participants. Long sleep was linked to higher mortality (relative risk 1.39, 95% CI 1.31 to 1.47), diabetes (1.26), cardiovascular disease (1.25), stroke (1.46), and coronary heart disease (1.24), with no significant link to hypertension (1.01).
Limits: Observational, and the authors say whether long sleep is a cause or a modifiable risk factor remains to be tested.
Yin J, Jin X, Shan Z, Li S, Huang H, Li P, et al. Relationship of Sleep Duration With All-Cause Mortality and Cardiovascular Events: A Systematic Review and Dose-Response Meta-Analysis of Prospective Cohort Studies. Journal of the American Heart Association. 2017;6(9):e005947.Meta-analysisPubMed 28889101 · DOI
What it found: Dose-response meta-analysis of prospective cohorts in generally healthy people. Every outcome showed a U-shaped curve with risk lowest at about 7 hours a day; for all-cause mortality the relative risk was 1.06 per hour below 7 hours and 1.13 per hour above, and for stroke 1.18 per extra hour.
Limits: Observational cohorts with self-reported sleep, so the curve shows association, not that changing your sleep changes your risk.
Kripke DF, Garfinkel L, Wingard DL, Klauber MR, Marler MR. Mortality associated with sleep duration and insomnia. Archives of General Psychiatry. 2002;59(2):131-136.Cohort studyPubMed 11825133 · DOI
What it found: American Cancer Society Cancer Prevention Study II: more than 1.1 million adults aged 30 to 102 reported their sleep. Those sleeping 7 hours had the best survival, mortality hazard was significantly higher for 8 hours or more and for 6 hours or less, and the authors wrote that short sleep and insomnia seem associated with little risk distinct from comorbidities and that causality is unproven.
Limits: Self-reported sleep in an older observational cohort from the 1980s, and the authors state that causality is unproven.
Itani O, Jike M, Watanabe N, Kaneita Y. Short sleep duration and health outcomes: a systematic review, meta-analysis, and meta-regression. Sleep Medicine. 2017;32:246-256.Meta-analysisPubMed 27743803 · DOI
What it found: Meta-analysis of 153 studies with 5,172,710 participants. Short sleep was linked to higher mortality (relative risk 1.12, 95% CI 1.08 to 1.16), diabetes (1.37), obesity (1.38), hypertension (1.17), cardiovascular disease (1.16), and coronary heart disease (1.26), and mortality rose in a linear pattern below six hours.
Limits: Observational pooling of self-reported sleep, so it shows associations and cannot separate short sleep from the illnesses, jobs, or habits that go with it.
Shi G, Xing L, Wu D, Bhattacharyya BJ, Jones CR, McMahon T, et al. A Rare Mutation of β1-Adrenergic Receptor Affects Sleep/Wake Behaviors. Neuron. 2019;103(6):1044-1055.e7.Lab or mechanistic studyPubMed 31473062 · DOI
What it found: The authors identified a mutation in the β1-adrenergic receptor gene (ADRB1) in people who need less sleep than most. The mutation reduces protein stability in cells, mice carrying it sleep for shorter periods, and the ADRB1-positive neurons in the dorsal pons are active in REM sleep and wakefulness.
Limits: A rare mutation in one family studied with cells and mice, so it does not explain most short sleepers.
He Y, Jones CR, Fujiki N, Xu Y, Guo B, Holder JL Jr, et al. The transcriptional repressor DEC2 regulates sleep length in mammals. Science. 2009;325(5942):866-870.Lab or mechanistic studyPubMed 19679812 · DOI
What it found: The authors identified a mutation in the human transcriptional repressor DEC2 that is linked to short sleep, and mice carrying the equivalent mutation slept less than controls, with the difference depending on time of day and on sleep deprivation.
Limits: A rare family-level mutation plus a mouse model, so it explains how a few people sleep less and says nothing about whether other people can train themselves to.
Yetish G, Kaplan H, Gurven M, Wood B, Pontzer H, Manger PR, et al. Natural sleep and its seasonal variations in three pre-industrial societies. Current Biology. 2015;25(21):2862-2868.Cross-sectional studyPubMed 26480842 · DOI
What it found: Sleep recorded in three pre-industrial societies (the record lists Bolivia, Namibia, and Tanzania). Sleep periods averaged 6.9 to 8.5 hours and actual sleep 5.7 to 7.1 hours, near the low end of industrial societies; sleep started about 3.3 hours after sunset, and naps occurred on fewer than 7% of winter days and fewer than 22% of summer days.
Limits: Three small groups of people living very differently from us, so it shows what unplugged sleep looks like, not what health outcomes those hours produce.
National Heart, Lung, and Blood Institute. Sleep Phases and Stages. NHLBI, National Institutes of Health, March 24, 2022.ArticleRead
What it found: NHLBI page stating that each sleep cycle lasts about 80 to 100 minutes, that most people go through four to six cycles a night and may wake briefly between them, and that deep sleep comes mostly early in the night and REM sleep mostly later.
Limits: A patient-education page giving an average, and cycle length varies between people and across the night.
Dickstein C. If you're tired take a nap, according to new Army field manual focused on holistic health. Stars and Stripes, October 1, 2020.ArticleRead
What it found: Stars and Stripes report on the Army's Holistic Health and Fitness field manual: soldiers should get 7 to 9 hours of sleep in every 24 hours, adapting to sleep loss is a myth, short infrequent naps can restore alertness when night sleep isn't possible, and with restricted sleep (six hours or less a night) napping plus appropriate doses of caffeine (about 200 mg, in the report's bracketed note) may help sustain performance.
Limits: A news report on the manual, not the manual itself, so wording should be checked against the field manual.
Watson NF, Badr MS, Belenky G, Bliwise DL, Buxton OM, Buysse D, 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.GuidelinePubMed 26039963 · DOI
What it found: The American Academy of Sleep Medicine and the Sleep Research Society used a modified RAND Appropriateness Method to produce a consensus recommendation on how much sleep healthy adults need for optimal health.
Limits: The PubMed abstract does not give the hour figures; the recommendation itself (at least seven hours a night) is in the full text, which I read on PubMed Central.
Tassi P, Muzet A. Sleep inertia. Sleep Medicine Reviews. 2000;4(4):341-353.ReviewPubMed 12531174 · DOI
What it found: Review of sleep inertia, the temporary drop in alertness after waking. Waking from slow-wave sleep causes the most inertia and waking from stage 1 or 2 the least, with REM in between; durations range from about 1 minute to 4 hours across studies but rarely exceed 30 minutes without major sleep deprivation.
Limits: A 2000 review that does not test alarms timed to 90-minute cycles, so it supports the idea that the sleep stage at waking matters but not any particular alarm schedule.
Cappuccio FP, Cooper D, D'Elia L, Strazzullo P, Miller MA. Sleep duration predicts cardiovascular outcomes: a systematic review and meta-analysis of prospective studies. European Heart Journal. 2011;32(12):1484-1492.Meta-analysisPubMed 21300732 · DOI
What it found: Meta-analysis of 15 prospective studies (24 cohort samples, 474,684 participants, followed 6.9 to 25 years, 16,067 events). Short sleep went with more coronary heart disease (relative risk 1.48) and stroke (1.15) but not total cardiovascular disease (1.03); long sleep went with more coronary heart disease (1.38), stroke (1.65), and total cardiovascular disease (1.41).
Limits: Sleep was self-reported by questionnaire and the studies were observational, so the abstract does not claim sleep duration causes these events.
Cappuccio FP, D'Elia L, Strazzullo P, Miller MA. Quantity and quality of sleep and incidence of type 2 diabetes: a systematic review and meta-analysis. Diabetes Care. 2010;33(2):414-420.Meta-analysisPubMed 19910503 · DOI
What it found: Meta-analysis of 10 studies (13 cohorts, 107,756 participants, 3,586 new cases of type 2 diabetes, follow-up 4.2 to 32 years). Short sleep (about 5 to 6 hours or less) had a relative risk of 1.28 and long sleep (more than 8 to 9 hours) 1.48; trouble staying asleep carried the highest risk at 1.84.
Limits: Observational cohorts with self-reported sleep and significant heterogeneity, so the associations may reflect other conditions rather than sleep itself.
He M, Deng X, Zhu Y, Huan L, Niu W. The relationship between sleep duration and all-cause mortality in the older people: an updated and dose-response meta-analysis. BMC Public Health. 2020;20(1):1179.Meta-analysisPubMed 32723316 · DOI
What it found: Meta-analysis of 28 prospective studies covering 95,259 older adults. Long sleep was tied to higher all-cause mortality (adjusted hazard ratio 1.24, 95% CI 1.16 to 1.33) while short sleep was only marginal (1.04, 95% CI 1.00 to 1.09); publication bias was likely for long sleep, though the trim-and-fill method strengthened that result.
Limits: Older adults only, self-reported sleep, and the authors list sex, region, survey method, age, and follow-up as sources of inconsistency between studies.
Youngstedt SD, Kripke DF. Long sleep and mortality: rationale for sleep restriction. Sleep Medicine Reviews. 2004;8(3):159-174.ReviewPubMed 15144959 · DOI
What it found: Review arguing that studies consistently link sleeping more than 8 hours to higher mortality. In the 1.1 million American Cancer Society respondents, sleeping longer than 7.5 hours accounted for roughly 5% of total mortality, and the excess persisted after adjusting for 32 possible confounders.
Limits: A narrative review by authors who argue for testing sleep restriction, and they say epidemiologic data cannot prove long sleep causes death.
Li J, Cao D, Huang Y, Chen Z, Wang R, Dong Q, et al. Sleep duration and health outcomes: an umbrella review. Sleep & Breathing = Schlaf & Atmung. 2022;26(3):1479-1501.Systematic reviewPubMed 34435311 · DOI
What it found: Umbrella review of 69 meta-analyses. In short sleepers each hour less per 24 hours went with a 3% to 11% higher risk of all-cause mortality, coronary heart disease, osteoporosis, stroke, and type 2 diabetes; in long sleepers each hour more went with a 7% to 17% higher risk of stroke mortality, coronary heart disease, stroke, and type 2 diabetes in adults.
Limits: A review of meta-analyses of mostly observational studies, so it inherits their confounding and cannot show cause.
Pellegrino R, Kavakli IH, Goel N, Cardinale CJ, Dinges DF, Kuna ST, et al. A novel BHLHE41 variant is associated with short sleep and resistance to sleep deprivation in humans. Sleep. 2014;37(8):1327-1336.Lab or mechanistic studyPubMed 25083013 · DOI
What it found: The authors sequenced BHLHE41 (also called DEC2) in two cohorts, 200 people with acute and 217 with chronic partial sleep deprivation, and found a Y362H variant in one twin of a non-identical pair, who slept less, had less recovery sleep, and had fewer performance lapses during deprivation than the twin without it. The abstract notes an earlier DEC2 mutation in a family of short sleepers who needed just 6 hours a night.
Limits: The key finding is a single twin pair plus cell work, so it shows such variants exist and says nothing about how common they are.
Walker M. Dr. Matthew Walker on Sleep for Enhancing Learning, Creativity, Immunity, and Glymphatic System. FoundMyFitness, February 28, 2019.PodcastListen
What it found: FoundMyFitness episode #45 with sleep scientist Matthew Walker of UC Berkeley. He recommends somewhere between seven and nine hours for adults, says short sleep predicts all-cause mortality, links six hours or less to higher risk of several cancers, estimates that cutting the last two hours of an eight-hour night can cost almost 80% of REM sleep, and argues hunter-gatherers are not a good control for how modern people should sleep.
Limits: A long podcast page in which Walker's claims, including the REM estimate, are his own statements, not a controlled finding.
Lieberman HR, Tharion WJ, Shukitt-Hale B, Speckman KL, Tulley R. Effects of caffeine, sleep loss, and stress on cognitive performance and mood during U.S. Navy SEAL training. Sea-Air-Land. Psychopharmacology. 2002;164(3):250-261.Randomized trialPubMed 12424548 · DOI
What it found: Sixty-eight Navy SEAL trainees were randomly given 100, 200, or 300 mg of caffeine or placebo after 72 hours of sleep deprivation with continuous stress. Sleep loss and environmental stress worsened performance and mood, and caffeine at 200 and 300 mg improved vigilance, reaction time, and alertness; the authors suggest 200 mg was the best dose.
Limits: A study of imposed sleep loss during SEAL training, which says nothing about a chosen nightly sleep schedule.
Rosekind MR, Smith RM, Miller DL, Co EL, Gregory KB, Webbon LL, et al. Alertness management: strategic naps in operational settings. Journal of Sleep Research. 1995;4(S2):62-66.ReviewPubMed 10607214 · DOI
What it found: NASA Ames review concluding that naps improve later performance and alertness in operational settings, that planned in-flight naps helped crews during descent and landing on a long-haul study, and that sleep inertia (grogginess just after waking) should be weighed before adding naps to schedules.
Limits: A short conference-style review from operational fatigue research that reports no specific numbers in its abstract.
Watson NF, Badr MS, Belenky G, Bliwise DL, Buxton OM, Buysse D, et al. Joint Consensus Statement of the American Academy of Sleep Medicine and Sleep Research Society on the Recommended Amount of Sleep for a Healthy Adult: Methodology and Discussion. Sleep. 2015;38(8):1161-1183.GuidelinePubMed 26194576 · DOI
What it found: Companion paper to the AASM and SRS consensus statement describing the background literature, the panel's voting process and results, key assumptions, and future research directions on sleep need and sleep health.
Limits: I read only the abstract of this long methods paper, which does not state the recommended hours.
Van Dongen HP, Maislin G, Mullington JM, Dinges DF. The cumulative cost of additional wakefulness: dose-response effects on neurobehavioral functions and sleep physiology from chronic sleep restriction and total sleep deprivation. Sleep. 2003;26(2):117-126.Randomized trialPubMed 12683469 · DOI
What it found: Healthy adults (n = 48, aged 21 to 38) were randomized to 4, 6, or 8 hours in bed for 14 nights, or to 3 nights without sleep. Restriction to 6 hours or less produced cognitive deficits that grew each day and matched up to 2 nights of total sleep deprivation, while subjective sleepiness ratings did not track the deficits.
Limits: A laboratory study of 48 healthy young adults over 14 days, so it shows what happens during restriction and not what happens over years.
Belenky G, Wesensten NJ, Thorne DR, Thomas ML, Sing HC, Redmond DP, et al. Patterns of performance degradation and restoration during sleep restriction and subsequent recovery: a sleep dose-response study. Journal of Sleep Research. 2003;12(1):1-12.Randomized trialPubMed 12603781 · DOI
What it found: Sixty-six healthy volunteers were assigned to 3, 5, 7, or 9 hours in bed for 7 days, then 8 hours for 3 recovery nights. Performance on a vigilance task kept falling in the 3-hour group, leveled off at a lower level in the 5- and 7-hour groups (with lapses rising only at 5 hours), and stayed at baseline at 9 hours; the 5- and 7-hour groups showed no meaningful recovery after 3 nights.
Limits: A one-week laboratory study of 66 volunteers, so it cannot show long-term effects or whether people adapt over months.
Spiegel K, Leproult R, Van Cauter E. Impact of sleep debt on metabolic and endocrine function. Lancet (London, England). 1999;354(9188):1435-1439.Lab or mechanistic studyPubMed 10543671 · DOI
What it found: Eleven young men spent 4 hours in bed a night for 6 nights and then 12 hours a night for 6 nights. During sleep debt, glucose tolerance was lower (p < 0.02), evening cortisol was higher (p = 0.0001), and sympathetic nervous system activity was higher (p < 0.02).
Limits: Eleven young men in a fixed-order lab study at 4 hours in bed, a more extreme restriction than 4.5 to 6 hours, over only 6 nights, with no randomization.
Kurina LM, McClintock MK, Chen JH, Waite LJ, Thisted RA, Lauderdale DS. Sleep duration and all-cause mortality: a critical review of measurement and associations. Annals of Epidemiology. 2013;23(6):361-370.Systematic reviewPubMed 23622956 · DOI
What it found: Systematic review of 42 prospective studies drawn from 35 distinct populations. Unlike earlier reviews, it found no consistent association between self-reported sleep duration and mortality, and it called a robust U-shaped link premature because survey-based sleep measures correlate poorly with objective ones.
Limits: A critical review that points to measurement error, response bias, confounding, and reverse causation, so it argues against over-reading these studies in either direction.
Ferrie JE, Shipley MJ, Cappuccio FP, Brunner E, Miller MA, Kumari M, et al. A prospective study of change in sleep duration: associations with mortality in the Whitehall II cohort. Sleep. 2007;30(12):1659-1666.Cohort studyPubMed 18246975 · DOI
What it found: Whitehall II cohort of 10,308 British civil servants aged 35 to 55 at baseline, followed 17 years. Among people sleeping 6, 7, or 8 hours, a decrease in sleep went with cardiovascular death (hazard ratio 2.4, 95% CI 1.4 to 4.1), and among those sleeping 7 or 8 hours, an increase went with non-cardiovascular death (2.1, 95% CI 1.4 to 3.1); adjustment for existing illness and health behaviors left these largely unchanged.
Limits: Observational, so it cannot show whether a change in sleep caused the deaths or reflected something else, such as developing illness.
Patrick R. Habits That Support Better Sleep. FoundMyFitness (clip from Rhonda Patrick's appearance on The Mel Robbins Podcast), July 16, 2026.PodcastListen
What it found: A FoundMyFitness clip in which Patrick suggests aiming for about seven and a half to nine hours of actual sleep, not time in bed, plus morning light, a steady wake time, no eating within about three hours of bed, and caution about alcohol near bedtime.
Limits: A short clip summarized on a show-notes page, so her reasoning and any caveats beyond the page text are not captured.
Patrick R. How Heat Therapy Improves Slow Wave Sleep. FoundMyFitness, May 30, 2023.PodcastListen
What it found: FoundMyFitness episode in which Patrick says heat therapy, such as a sauna or a hot bath, may improve slow-wave sleep; the example protocols are 20 minutes in a 176°F (80°C) sauna or a 104.9°F (40.5°C) hot tub or bath for 20 to 30 minutes, 1 to 2 hours before bedtime.
Limits: The page names no studies for its sleep claim, and the numbers are the host's description of the literature, not a trial result.