Nap Length Predicts the Cost Better Than the Benefit

Two forty in the afternoon. You have read the same paragraph of the spec three times and the third reading went worse than the second. You know the rule, because everyone knows the rule: twenty minutes. So you set a timer for twenty minutes, lie down on the floor of the spare room or tip the desk chair back, and spend the first eleven of those minutes not sleeping and thinking about the timer. The alarm goes off. You got maybe nine minutes. You feel roughly the way you felt before, except now it is three o’clock.

Most people conclude at that point that they are bad at napping. Here is what the best-instrumented trial on the question found: to obtain ten minutes of actual sleep, verified by polysomnography, its participants needed an average of 24.5 minutes in bed [1]. The timer was measuring the wrong thing.

Napping belongs on a site about working with your head because it is a scheduling decision you make inside a working day, and because the trials at the centre of it measure what we care about — processing speed, vigilance, memory encoding — in adults aged 21 to 35, in the early afternoon [1]. The connection is direct. The size of the effect is smaller and stranger than you have been told.

I went looking for the optimal number. What I found instead: the number everybody repeats is a composite of three unrelated measurements, and the paper credited with the first of them does not contain it anywhere I can read. A peer-reviewed review went looking for the rule’s evidence base in 2017 and reported that the picture was unclear. A regulator’s systematic review reached close to the opposite conclusion six years later, and I cannot make the two agree. And underneath all of it sits an asymmetry no popular source states — nap length is a far better predictor of what a nap costs you than of what it gives you.

Three numbers, none of which was ever an instruction

Thirty is supposed to come from physiology: the typical delay before slow-wave sleep begins. The sources that do attribute the figure point to Gillberg and Åkerstedt, 1991.

I went and looked that paper up [2]. It is called The Dynamics of the First Sleep Cycle. It ran eight subjects. The design was four conditions of progressively postponed night sleep — eight, four, two or zero hours — each followed by a daytime sleep beginning at eleven in the morning. Its reported findings were that the duration of the first sleep cycle did not change significantly across conditions, and that as prior sleep loss increased, slow-wave energy and slow-wave activity increased with it.

I have that paper’s abstract and indexed record, not its full text, and no thirty-minute figure appears in either. The number enters circulation downstream instead — a 2025 paper states that in normal nocturnal sleep deep sleep begins roughly thirty minutes after sleep onset, and cites Gillberg and Åkerstedt for it [4].

The study measured how the shape of the first sleep cycle changes with accumulated sleep debt. It did not test what happens when you end a sleep period at a chosen minute, because nobody in 1991 was asking that. The rule needs a cut-off point. The paper measured a curve.

And the finding the paper does report cuts against the rule built on it. How much deep sleep arrives, and how fast, tracked prior sleep debt. Your Tuesday and your Thursday are different experiments.

Twenty-six comes from an aircraft. In the early 1990s NASA gave 21 long-haul pilots a planned 40-minute rest opportunity during low-workload cruise, with EEG running [3]. They fell asleep on 93% of opportunities, took about 5.6 minutes to drop off, and slept an average of 25.8 minutes. Post-nap reaction time improved.

The Sleep Foundation’s page on this states that NASA determined the ideal nap length to be 26 minutes. NASA determined no such thing. NASA offered forty minutes and measured what happened inside them. The operational rule that came out of aviation guidance is a 40-minute cap on the opportunity — and a later analysis notes that even in the original study, the 40-minute cap did not prevent deep sleep [4]. An observation about how long people slept became an instruction about how long to sleep, and the conversion silently deleted the time it takes to fall asleep.

Two percentages also travel with this study, describing improvements in performance and in physiological alertness. They come from a 1995 conference-supplement paper rather than the technical memorandum, and I could not verify either against a primary document. They are not evidence in this piece.

Ninety is the mean length of a sleep cycle. It was never measured in a nap-duration trial.

Fuse the three and you get the rule as it appears, in near-identical wording, on Calm, Whoop, Amerisleep, Healthline and a hundred content sites: nap under thirty minutes, or go a full ninety to complete a cycle, never in between. I could not find a source for it on any page that states it. It is attributed to nobody.

The unit problem, which is the part worth learning

Before the trial evidence makes sense, notice that “a twenty-minute nap” means at least three incompatible things.

Some trials define the nap as minutes of sleep, verified by EEG. Others define it as time until awakening, whatever happened in there. NASA defined it as opportunity offered. Popular coverage flattens all three into one phrase. That is the difference between a thirty-minute nap containing thirty minutes of sleep and a thirty-minute nap containing nine.

Which is why you should be suspicious of any article that lays nap durations on a single axis and draws a curve. Here are the trials that compared more than one nap length in the ten-to-sixty-minute range people actually schedule:

Different lengths, different definitions, different outcomes, different people — one run at four in the morning, one in men only who had been awake for twenty to thirty hours. No trial anywhere compares 10, 20, 60 and 90 minutes. The dose-response curve you have seen was assembled, not measured.

The one part of the rule that holds at the bottom: naps of 30 and 90 seconds produced no measurable improvement on anything, while a 10-minute nap did [10]. There is a real floor, somewhere between ninety seconds and ten minutes.

Somebody did check, in 2017

For three drafts of this piece I believed nobody had audited the rule. That was wrong, and the correction improves the story.

In 2017, Hilditch, Dorrian and Banks published a review in Sleep Medicine whose framing is almost word for word the one I arrived at independently: many industry publications recommend naps of thirty minutes or less to avoid sleep inertia, and the evidence supporting that advice had not been thoroughly reviewed [11]. So they reviewed it.

Their finding: studies show mixed results on both when slow-wave sleep begins and how long and severe the inertia is afterwards, which leaves the guidelines unclear. The variation, they suggest, comes from differences in what people slept the night before and what time of day they woke. They called for guidance specified by time of day and prior sleep history rather than by duration alone.

The audit exists. It has been in a peer-reviewed journal since 2017. What does not exist is any sign that the advice built on the rule was updated afterwards.

What the EEG saw

In 2023, Leong and colleagues ran the cleanest version of the experiment: 32 habitual short sleepers, every person doing every condition, all three nap lengths verified by polysomnography, testing at 5, 30, 60 and 240 minutes afterwards [1].

The stage people were in when the alarm went off:

• 10-minute nap — 6.3% woke out of deep sleep
• 30-minute nap — 54.8% woke out of deep sleep
• 60-minute nap — 6.7% woke out of deep sleep

The thirty-minute nap is the one the rule declares safe, on the grounds that it stops short of deep sleep. In this trial it was the length that landed in deep sleep most often — more than half the time — while the sixty-minute nap had mostly carried people through and out the other side. Twenty per cent of the sixty-minute nappers woke from REM. The rule aims at avoiding N3 and lands on it.

The mechanism underneath is quantified elsewhere. In a night-shift protocol, a 30-minute nap contained 14.7 ± 5.7 minutes of slow-wave sleep and a 10-minute nap contained 0.8 ± 1.5 minutes [9] — roughly eighteen times as much, arriving in the twenty extra minutes. Two caveats travel with that. It was run at four in the morning, so it cannot be moved to a two o’clock desk nap unaltered. And in that study neither nap improved performance, at any point in the inertia window or in the two and a half hours that followed [12]. A paper titled for a difference in cost found no difference in benefit, because there was no benefit to differ.

Two careful reviews, and I cannot reconcile them

I expected one of them to be wrong. Neither is obviously wrong, and reporting the disagreement is more honest than picking a winner.

In December 2023, the European Union Aviation Safety Agency commissioned a systematic review to settle a question a regulator needed settled: is the thirty-minute sleep-inertia assumption it relies on valid? The resulting report screened 1,413 abstracts with two independent reviewers and extracted 70 studies [13]. Its verdict on nap length is the strongest evidence for a duration rule anywhere in this literature: reading four direct-comparison trials, and independent of whether the nap fell in the day or the night, naps of twenty minutes or less consistently showed no inertia effects on performance, while naps of thirty and sixty minutes produced decrements on waking. Across 21 studies comparing napping with staying awake, 14 found significant impairment after waking, and those impairments were mostly found after naps of thirty minutes or longer.

That is a duration effect, found by aerospace researchers with no stake in anyone’s blog post. It sits against the 2017 review’s “mixed and unclear.”

It also sits against a third result. Signal and colleagues, testing 20-, 40- and 60-minute nap opportunities with polysomnography, reported that no more than fifteen minutes was needed for inertia-related performance decrements to dissipate after any nap opportunity of sixty minutes or less — and, in a separate analysis, no effect of slow-wave sleep duration on inertia at all [7]. Their conclusion was that under conditions of short sleep, slow-wave sleep by itself does not make inertia worse. That is the mechanism the entire thirty-minute rule assumes, and it is contested. On one side, a 1985 study found slow-wave sleep to be the major driver of inertia [14] — a paper I have only through the EASA review, not from the original. On the other, an earlier simulated-night-shift study by Lovato and colleagues, which I likewise have only through a later review, reported that 91% of participants woke out of slow-wave sleep with no impairment at all [15]. The best available summary is that a clear conclusion cannot be synthesised [12].

Worth noting plainly: Signal 2012 is a direct duration comparison with PSG, and it is not among the four studies EASA drew its duration conclusion from. Whether it sat elsewhere among the 70 extracted is not something I can check without the report’s data annex.

One caution about the architecture story, since this piece leans on it more than once. In the 2023 afternoon trial, no sleep-stage measure was significantly associated with any outcome, and the authors could not attribute the benefits they found to architecture [1]. Architecture beat duration in the driving study. In the best-instrumented afternoon trial, architecture predicted nothing at all — while duration went on predicting the cost of waking and not the size of the benefit.

Where does this leave the most-repeated consumer advice? Andrew Huberman’s published protocol recommends keeping naps to twenty minutes or less to avoid sleep inertia, and adds that if you tend to feel groggy afterwards you can skip naps entirely, since they are not necessary. That lands on the side EASA supports. The narrower claim that twenty minutes avoids inertia is the one the 2017 review examined and called unclear, and one 2017 study found impairment immediately after a ten-minute nap taken at 07:00 following extended wakefulness [16]. The advice is better supported than its sourcing suggests, and narrower than its phrasing suggests.

The half of the equation nobody separates out

Here is the asymmetry. On cost, duration mostly works — more minutes in the nap means more deep sleep in it, and the impairment on waking is larger and better documented at thirty and sixty minutes than at ten. What duration does not buy you is control over which stage the alarm catches. On benefit, duration has been looked for repeatedly and mostly not found.

• A meta-analysis of 11 studies in working-aged adults reported that nap duration did not influence cognitive performance [17] — a paper whose first author’s affiliation string includes an occupational-health software company, not declared as a conflict. Its own pooled naps averaged 55.4 ± 29.4 minutes and ranged from 15 to 90, so this is a null across a range far wider than the one the rule argues about.
• A meta-analysis pooling 60 samples tested five moderators — age, nap length, nap start time, habitual napping, and prior sleep restriction — and found no significant moderation by any of them (Q values 0.009 to 8.572, all p > 0.116) [18]. That meta-analysis came from the same laboratory as the 2023 trial, so the two are not fully independent supports.
• A 2025 driving-simulator trial found duration predicted neither sleepiness nor driving errors, while sleep architecture predicted both [4] — run at a commercial traffic-sciences institute, with no funder stated in the record.

Now the boundary condition, because it is real and it is the strongest counter-evidence I found. In the Signal and Mulrine protocols, participants stayed awake for twenty to thirty hours before napping. Under those conditions duration did buy benefit: the slowest ten per cent of vigilance responses were significantly faster after 40- and 60-minute naps than after a 20-minute nap or none, there were fewer lapses after 60 minutes than after 20, and after 60 minutes people felt less sleepy and made more correct responses on a working-memory task [8]. Note when that benefit arrives: roughly forty-five minutes after waking, on the far side of the inertia window rather than inside it [7].

That is a dose-response in the benefit, on polysomnography, in a within-subjects design. It also describes two groups of twelve young men, awake for about twenty hours in one protocol and about thirty in the other. When you are genuinely sleep-deprived, more sleep helps more — the least surprising finding in sleep science, and not what the thirty-minute rule is about.

So the honest statement has a boundary in it. In the afternoon-nap literature, in people who are not severely sleep-deprived, duration has repeatedly failed to predict how much you get. Under heavy sleep debt it predicts it fine. Neither of those is the claim the rule makes, which is that duration is the dial setting your payoff on an ordinary Tuesday.

Be careful with the stack of nulls, too. Between-study moderation is a blunt instrument, the underlying trials are small, and the working-aged-adults meta-analysis found its eleven studies reported at roughly 42% of the CONSORT standard [17]. What survives is narrower than “duration does not matter” and still useful: people have gone looking for a dose-response in the benefit, more than once, with the best tools available, and outside deprivation studies they have not found one.

There is one place where a null is a clean answer rather than a failure to detect. In the 2023 trial, processing speed showed no benefit at any duration and any interval — accuracy on the test was null throughout, all p-values above 0.14, and the speed benefit was not observed at all — in a within-participant design where everyone did every condition [1]. Not a hint of an effect awaiting a bigger sample. An answer — and it is the answer to the thing most people want from a nap.

Which makes the following worth sitting with. The 2006 trial found the opposite on the same instrument: after a 10-minute nap, correct responses on the digit-symbol substitution test improved against rest at 35, 95 and 155 minutes post-nap — a breakdown I have through the 2025 trial’s discussion of it, not from the 2006 paper itself [5]. Two of the best trials in the field, one test, opposite results. Nobody has reconciled them.

How big is the benefit overall? Three meta-analyses from the last five years put the pooled effect of napping on cognition at 0.18 [17], 0.379 [18] and 0.69 [19]. A factor of nearly four. They are not pooling identical study sets or identical constructs, but no reader would guess that from the way any of them is cited, and every popular article picks exactly one. If you want a single number, 0.379 with the range acknowledged is the fairest available. None of them describes a transformed working afternoon.

You will not be able to tell whether it worked

This is the part that should change how you think about the question.

In 2025, a randomised trial put 81 healthy adults, average age 33.6, into three arms: a nap ended by a technologist watching live EEG, a nap ended by a wearable sensor, and a control group that sat upright watching a film [20]. The manual group slept about 11.8 minutes and 7.4% reached deep sleep. The sensor group slept about 22.6 minutes and 63.0% reached deep sleep.

The longer sleepers reported less sleepiness and less fatigue than controls in every post-nap session, and their objective sleepiness marker was lower in every session too. Their cognitive performance was statistically indistinguishable from people who never napped — the study’s primary analysis, null across all six sessions. The shorter sleepers showed the reverse: a cognitive improvement, in one session of six, on a secondary analysis the authors say was not corrected for multiplicity and should be read cautiously.

The nap that felt better worked worse. One trial, non-blinded, in people who were not sleep-deprived — but pointed the same way as a much larger body. Across 30 studies measuring both, the EASA review found the majority reported a disconnect between self-report and performance, with no consistent direction: 11 found subjective impairment outlasting objective, 12 found the reverse, 7 found no difference. Inside those 12, nine reported people showing measurable impairment while reporting no sleepiness or fatigue at all [13]. Signal’s group put it more bluntly: subjective sleepiness is not a reliable indicator of inertia [7].

There is no self-report you can run on this.

Two footnotes on that 2025 trial, because they are the pattern of the whole topic in miniature. The study was supported by KYOCERA Corporation; a co-author is a KYOCERA employee and developed the awakening system the study evaluates [20]. And the “wake nine minutes after stage 2” protocol the device implements traces to a one-page conference abstract from 2014 [21], not a peer-reviewed trial. The device used in the trial detected its own target moment 25.9% of the time. A redesigned second-generation device, tested separately, reached 36.0% within a ±6-minute window, and 80% only when the window was widened to ±16 minutes [20].

One more, recent enough that you may have seen the headlines. A 2023 Mendelian randomisation study reported that genetically predicted daytime napping was associated with larger total brain volume — β = 15.80 cm³, with a confidence interval running from 0.25 to 31.34 [22]. Press coverage turned this into naps keeping the brain younger by up to 6.5 years, a figure that came from an interview rather than the paper’s results. The paper also reports a weighted-median sensitivity estimate that did not reach significance, and null results for hippocampal volume, reaction time and visual memory. The authors’ own summary in the university release was that the study points to a causal link between habitual napping and larger total brain volume, and that they hoped work showing the health benefits of short naps would reduce the stigma around napping. Average participant age was 57. And the exposure was a three-option survey question — never/rarely, sometimes, usually — containing no information about nap length, instrumented by variants explaining 1% of the variance in napping. It is cited in nap-length coverage anyway.

What this can actually cost you

The acute cost is decisions, not health. Impairment in the first fifteen to thirty minutes after waking can be comparable to the effects of substantial sleep loss [12]. What inertia takes is specific: in a study of twenty junior officer reservists woken abruptly at 03:00 after under three hours of sleep and handed a tactical planning exercise with a mid-task change of situation, eight of ten failed the task versus three of ten controls [23]. The domains that broke were judgement — extracting relevant from irrelevant information, assessing available assets and cover. Trained procedural skills were unaffected. That design stacks short sleep, the circadian trough and inertia together; it is not a study of nap length and cannot describe a two o’clock nap. Its value is qualitative: inertia takes your ability to handle something changing, not your ability to execute a routine.

Short naps are not automatically free, either. A preliminary study of nine people taking 20-minute naps every six hours across 64 hours of sleep deprivation reported an extreme form of inertia, in which waking during the circadian temperature trough became distressing enough that the authors describe it as producing an aversion to napping [24]. Nine participants, 1993, and I could not obtain the full text. Take it as a flag, not a finding.

The time cost is real and systematically understated. Obtaining 60 minutes of sleep took 76.3 minutes in bed [1]. Add roughly 30 minutes of degraded output afterwards and a “one-hour nap” is about 106 minutes of your working day.

The epidemiology is about a different person than you. Habitual naps of an hour or more are associated with higher cardiovascular risk (RR 1.82, 95% CI 1.22–2.71) and higher all-cause mortality (RR 1.27, 1.11–1.45) in pooled cohorts of 151,588 people [25]. Naps under 60 minutes: null on both, p = 0.98 and p = 0.08. A 2024 meta-analysis of 21 cohorts put napping under an hour at a hazard ratio of exactly 1.00 (0.90–1.11, p = 0.971), though heterogeneity in that stratum was high, I² = 62.6%, with no significant association with cardiovascular disease either, while napping for an hour or more carried a hazard ratio of 1.22 for mortality and 1.37 for cardiovascular disease [26]. That paper reports its own sample size two ways — 371,306 in the abstract, 374,306 in the results — so treat the total as approximate.

Two limitations travel with all of it. These are observational cohorts and cannot establish direction; long habitual daytime napping is a known symptom of undiagnosed sleep apnoea, depression and subclinical cardiovascular disease, and a hazard ratio of 1.82 in a predominantly older cohort is at least as consistent with sick people napping longer as with long naps making people sick. And an umbrella review of 11 meta-analyses plus 97 articles found these associations mainly supported by weak or suggestive evidence, with napping under thirty minutes showing no significant risks and no association with cardiovascular risk in young and middle-aged adults — the significant dose-response appeared in adults over 60, for any daytime napping [27]. That review also notes primary studies usually categorise napping as yes/no or split only at one hour, so the effect of naps under thirty minutes mostly was not measured at all.

None of this is a finding about a 32-year-old taking a nap on a bad Tuesday.

One honest unknown: nobody has run a trial of habitual napping as a sustained practice with cognitive endpoints. Every trial above is a single afternoon, and the longest follow-up in the literature is four hours [1].

What to change on Monday, in order of how well it is supported

1. Book the calendar, not the sleep. The best-evidenced item here and the one nobody tells you. Ten minutes of sleep took 24.5 minutes in bed; thirty minutes took 43.8 [1]. Sleep latency in that trial had a standard deviation of about seven minutes — one person is out in three minutes, another in seventeen, and the same person differs day to day. Block 40 to 45 minutes if you want thirty minutes of sleep, 25 if you want ten, and stop setting a twenty-minute timer as though it were a dose.

2. Put nothing requiring judgement in the half hour after a nap of thirty minutes or more. The processing-speed cost after the 30- and 60-minute naps was significant at 5 minutes post-waking and gone by 30 minutes [1]. A scheduling constraint, not a reason to skip the nap. No deploys, no customer calls, no design reviews, no driving.

3. Drink the coffee before, not after. In an afternoon driving-simulator study, 200 mg of caffeine taken immediately before a nap of under fifteen minutes cut incidents to 9% of placebo levels, against 34% for the same dose of caffeine without the nap [28]. Caffeine taken before a sleep opportunity has been reported to eliminate psychomotor deficits from inertia [29]; caffeine taken after waking truncates inertia without touching the worst initial window [30]. Two caveats: the task was monotonous driving, not knowledge work, and that trial compared the combination against caffeine alone and placebo — it had no nap-only arm, so it does not show the combination beating a nap by itself. An earlier study from the same lab tested a nap and 150 mg of caffeine separately and found each reduced driving impairment on its own [31]. Most of us already drink the coffee. We drink it at the wrong end.

4. Pick the length by what you want, and hold it loosely. For alertness over the next hour, ten minutes was sufficient and produced no measurable inertia [1][5][13]. The sixty-minute nap was the opposite trade: null for vigilance response times and null for lapses at every interval in the 2023 trial, while producing the largest processing-speed cost on waking. One caveat on those vigilance nulls — no participant logged more than two lapses in any session, so the measure may have had little room to move. For encoding something you learned that afternoon, thirty minutes was the only length producing a positive result — one trial, 32 people [1]. If you have been badly short of sleep rather than mildly, the calculus flips and longer wins [8]. Going past thirty minutes on an ordinary day: sixty produced learning gains only when the nap contained REM sleep, and without REM a sixty-minute nap produced no improvement at all (P = 0.72) [32]. A sleep cycle is not a schedulable object.

5. Stop counting a failure to fall asleep as a failed nap. In the driving study, naps consisting of non-sleep dozing were still effective [28]. Lying still in a dim room with your eyes shut is not nothing, and treating it as nothing is how people talk themselves out of the practice entirely.

What you should not do is try to work out whether it worked. Thirty studies say the relationship between how you feel and how you perform is inconsistent in direction, and in nine of them people were measurably impaired while reporting themselves fine [13]. The mood and alertness gain is the most reliable thing in this literature across groups of people — still present at the four-hour mark, which is as far as the 2023 trial followed anyone. That is not the same as being able to read it off your own afternoon, and in the 2025 trial those subjective gains came from an unblinded design, which is exactly where expectation shows up. So: schedule around the cost, expect the mood effect on the average of your Tuesdays rather than on any one of them, and treat the cognitive benefit as a small bonus that may or may not have arrived.

About this article

Written by Drew Anton. Drew covers behaviour, sleep, and the compounds people take for both — doses, timing, and what the label does and does not tell you. Not a physician or research scientist — reads the primary literature closely and refuses to round up.

Medical review: None. NeuriFuel does not currently have a licensed clinician on the editorial team, and this article has not been medically reviewed. We state this rather than implying an authority we do not have. See our About page for our full methodology.

Sources: 33 documents across 32 numbered entries — entry 3 holds two NASA papers. Two systematic reviews (one peer-reviewed, one commissioned by a regulator and not peer-reviewed in the journal sense); six meta-analyses and umbrella reviews; one narrative review; twenty-two human trials and field studies; one Mendelian randomisation study; one conference abstract, flagged as such where it is used. Three sources were read in full: the 2023 nap-duration trial, the 2019 inertia review and the 2025 awakening trial. The rest are held at abstract and indexed-record level, and six only through the summaries in other reviews — flagged in the body wherever the claim depends on it, and marked for all six in the reference list. Funding and conflict-of-interest disclosures could not be retrieved for a substantial minority of the studies cited here — including the 2017 systematic review, the 2022 meta-analysis, the 2016 night-nap trial, the two 2012 nap-duration papers, and three of the four safety syntheses. That silence is not evidence the disclosures were clean; it means we could not read them. Where they were retrievable, they are stated inline. Values we could not verify are marked as unverified in the text, and conflicts between sources are flagged rather than resolved.

Corrections: Found an error? Write to hello@neurifuel.com with a source and we will fix it and log the correction.

Last updated: 13 August 2026

References

1. Leong RLF, Lau T, Dicom AR, Teo TB, Ong JL, Chee MWL (2023). Influence of mid-afternoon nap duration and sleep parameters on memory encoding, mood, processing speed, and vigilance. Sleep 46(4):zsad025. DOI 10.1093/sleep/zsad025. PMID 36775965. NCT04984824. Funded by the National Medical Research Council Singapore, the National Research Foundation, Yong Loo Lin School of Medicine and the Lee Foundation; authors declared no financial conflicts. Same laboratory as reference 18.
2. Gillberg M, Åkerstedt T (1991). The dynamics of the first sleep cycle. Sleep 14(2):147–154. DOI 10.1093/sleep/14.2.147. PMID 1866528. N = 8. Funding not retrieved. Held at abstract and indexed-record level; full text not read.
3. Rosekind MR, Graeber RC, Dinges DF, Connell LJ, Rountree MS, Spinweber CL, Gillen KA (1994). Crew Factors in Flight Operations 9: Effects of Planned Cockpit Rest on Crew Performance and Alertness in Long-Haul Operations. NASA Technical Memorandum; TM number not confirmed. Summary paper: Rosekind MR et al. (1995), Journal of Sleep Research 4(Suppl 2):62–66. NASA-funded. Conflicts not specified.
4. Tomzig M, Wörle J, Gary S, Baumann M, Neukum A (2025). Strategic naps in automated driving — sleep architecture predicts sleep inertia better than nap duration. Accident Analysis & Prevention 209:107811. DOI 10.1016/j.aap.2024.107811. PMID 39427445. Authors declared no competing financial interests. Lead affiliation WIVW GmbH, a commercial traffic-sciences institute; funder not stated in the retrieved record.
5. Brooks A, Lack L (2006). A brief afternoon nap following nocturnal sleep restriction: which nap duration is most recuperative? Sleep 29(6):831–840. DOI 10.1093/sleep/29.6.831. PMID 16796222. Not industry-supported; authors declared no financial conflicts. The 35/95/155-minute breakdown cited here is retrieved secondarily via reference 20, not from this paper.
6. Tietzel AJ, Lack LC (2001). The short-term benefits of brief and long naps following nocturnal sleep restriction. Sleep 24(3):293–300. DOI 10.1093/sleep/24.3.293. PMID 11322712. Funding not retrieved.
7. Signal TL, van den Berg MJ, Mulrine HM, Gander PH (2012). Duration of sleep inertia after napping during simulated night work and in extended operations. Chronobiology International 29(6):769–779. DOI 10.3109/07420528.2012.686547. PMID 22734577. Sleep/Wake Research Centre, Massey University, New Zealand. Funding not retrieved.
8. Mulrine HM, Signal TL, van den Berg MJ, Gander PH (2012). Post-sleep inertia performance benefits of longer naps in simulated nightwork and extended operations. Chronobiology International 29(9):1249–1257. PMID 23002951. Companion paper to reference 7; two protocols of 12 participants each, healthy young men only. Funding not retrieved.
9. Hilditch CJ, Centofanti SA, Dorrian J, Banks S (2016). A 30-minute, but not a 10-minute nighttime nap is associated with sleep inertia. Sleep 39(3):675–685. DOI 10.5665/sleep.5550. PMID 26715234. Funding and conflicts not retrieved. The performance result cited here is retrieved secondarily via reference 12.
10. Tietzel AJ, Lack LC (2002). The recuperative value of brief and ultra-brief naps on alertness and cognitive performance. Journal of Sleep Research 11(3):213–218. DOI 10.1046/j.1365-2869.2002.00299.x. PMID 12220317. Funding not retrieved.
11. Hilditch CJ, Dorrian J, Banks S (2017). A review of short naps and sleep inertia: do naps of 30 min or less really avoid sleep inertia and slow-wave sleep? Sleep Medicine 32:176–190. DOI 10.1016/j.sleep.2016.12.016. PMID 28366332. Frequently miscited as Sleep Medicine Reviews. Funding not retrieved.
12. Hilditch CJ, McHill AW (2019). Sleep inertia: current insights. Nature and Science of Sleep 11:155–165. DOI 10.2147/NSS.S188911. Hilditch supported by a NASA System Wide Safety grant; McHill by NIH K01HL146992 and the Oregon Institute of Occupational Health Sciences. McHill reports speaker honoraria or travel reimbursement from the Utah Sleep Research Society and the California Precast Concrete Association.
13. Fischer D, Elmenhorst E-M, Benderoth S, van Drongelen A (2023). Report D-4: Duration of Sleep Inertia. Deliverable under EASA contract EASA.2022.C17, 15 December 2023. Authors at DLR and NLR. Funded by the EU Horizon Europe Programme. A contracted deliverable, not peer-reviewed in the journal sense; carries an explicit disclaimer that it is not binding on EASA. No individual author conflict statement retrieved.
14. Dinges DF, Orne MT, Orne EC (1985). Behavior Research Methods, Instruments, & Computers 17(1):37–45. DOI 10.3758/BF03200895. Title not retrieved. Retrieved secondarily via reference 13; primary not opened.

15. Lovato N, Lack L, Ferguson S, Tremaine R (2009). Sleep and Biological Rhythms 7(1):34–42. Title not retrieved. Retrieved secondarily via reference 12. N, statistics and funding not retrieved.
16. Hilditch CJ, Dorrian J, Centofanti SA, Van Dongen HP, Banks S (2017). Sleep inertia associated with a 10-min nap before the commute home following a night shift. Accident Analysis & Prevention 99(Pt B):411–415. DOI 10.1016/j.aap.2015.11.010. PMID 26589387.
17. Dutheil F, Danini B, Bagheri R, Fantini ML, Pereira B, Moustafa F, Trousselard M, Navel V (2021). Effects of a short daytime nap on the cognitive performance: a systematic review and meta-analysis. IJERPH 18(19):10212. DOI 10.3390/ijerph181910212. PMID 34639511. States it received no external funding and that no conflicting relationship exists for any author. The first author’s affiliation string includes WittyFit, an occupational-health software company; this is not declared as a conflict and is recorded here as a fact.
18. Leong RLF, Lo JC, Chee MWL (2022). Systematic review and meta-analyses on the effects of afternoon napping on cognition. Sleep Medicine Reviews 65:101666. DOI 10.1016/j.smrv.2022.101666. PMID 36041284. Funding not retrieved. Same laboratory as reference 1.
19. Du P, Li J, Hua Z, Sun Y, Song S, Liao Y, Cheng S, Li X (2026). Public Health Reviews 47:1609013. DOI 10.3389/phrs.2026.1609013. Title and funding not retrieved.
20. Suzuki Y, Suzuki C, Suzuki Y, Kawana F, Ohigashi T, Maruo K, Watanabe T, Abe T (2025). Effects of optimal timed automatic awakening from a short daytime nap on cognitive performance, alertness, and fatigue. Scientific Reports 15:37228. DOI 10.1038/s41598-025-21008-3. PMID 41136428. UMIN000045658, UMIN000052386. Non-blinded. Industry-supported: funded by KYOCERA Corporation alongside NEDO, AMED and JSPS; one co-author is a KYOCERA employee and developed the awakening system evaluated in the study. Remaining authors declared no competing interests.
21. Hayashi M, Fushimi A, Iizuka H (2014). Nine minutes of sleep stage 2 is optimum for obtaining the greatest benefits from a daytime nap. International Journal of Psychophysiology 94:225. DOI 10.1016/j.ijpsycho.2014.08.886. Conference abstract, single page, not peer-reviewed. N, statistics and funding not retrieved.
22. Paz V, Dashti HS, Garfield V (2023). Is there an association between daytime napping, cognitive function, and brain volume? A Mendelian randomization study in the UK Biobank. Sleep Health 9(5):786–793. DOI 10.1016/j.sleh.2023.05.002. Funding and conflicts not retrieved. Instrument from Dashti et al. 2021, in which 123 variants explain 1% of the variance in daytime napping; 92 were used here after linkage-disequilibrium clumping.
23. Horne J, Moseley R (2011). Sudden early-morning awakening impairs immediate tactical planning in a changing ’emergency’ scenario. Journal of Sleep Research 20(2):275–278. DOI 10.1111/j.1365-2869.2010.00904.x. PMID 21518064. N = 20, ten per group. Funding not retrieved.
24. Naitoh P, Kelly T, Babkoff H (1993). Sleep inertia: best time not to wake up? Chronobiology International 10(2):109–118. DOI 10.1080/07420529309059699. PMID 8500187. Described by its authors as a preliminary study; nine participants in the nap condition and ten in the sleep-deprivation condition. Supported by the Naval Medical Research and Development Command, Department of the Navy, Work Unit 63706N MOO96 002.6002. Full text not obtained.
25. Yamada T, Hara K, Shojima N, Yamauchi T, Kadowaki T (2015). Daytime napping and the risk of cardiovascular disease and all-cause mortality. Sleep 38(12):1945–1953. DOI 10.5665/sleep.5246. PMID 26158892. Not industry-supported; authors indicated no financial conflicts.
26. Wang M, Xiang X, Zhao Z, Liu Y, Cao Y, Guo W, Hou L, Jiang Q (2024). Association between self-reported napping and risk of cardiovascular disease and all-cause mortality: a meta-analysis of cohort studies. PLOS ONE 19(10):e0311266. DOI 10.1371/journal.pone.0311266. The paper reports its pooled sample as 371,306 in the abstract and 374,306 in the results section. Funding not retrieved.
27. Sun J, Ma C, Zhao M, Magnussen CG, Xi B (2022). Daytime napping and cardiovascular risk factors, cardiovascular disease, and mortality: a systematic review. Sleep Medicine Reviews 65:101682. DOI 10.1016/j.smrv.2022.101682. Funding not retrieved.
28. Reyner LA, Horne JA (1997). Suppression of sleepiness in drivers: combination of caffeine with a short nap. Psychophysiology 34(6):721–725. DOI 10.1111/j.1469-8986.1997.tb02148.x. PMID 9401427. N = 12. Conditions were caffeine plus nap, 200 mg caffeine alone, and placebo; there was no nap-only condition. Listed by PubMed as receiving non-US government research support; funder not named.
29. Van Dongen HPA, Price NJ, Mullington JM, Szuba MP, Kapoor SC, Dinges DF (2001). Sleep 24(7):813–819. DOI 10.1093/sleep/24.7.813. Title not retrieved. Retrieved secondarily via reference 12.
30. Newman RA, Kamimori GH, Wesensten NJ, Picchioni D, Balkin TJ (2013). Perceptual and Motor Skills 116(1):280–293. DOI 10.2466/29.22.25.PMS.116.1.280-293. Title not retrieved. Retrieved secondarily via reference 12.
31. Horne JA, Reyner LA (1996). Counteracting driver sleepiness: effects of napping, caffeine, and placebo. Psychophysiology 33(3):306–309. DOI 10.1111/j.1469-8986.1996.tb00428.x. PMID 8936399. N = 10, 150 mg caffeine. Funding not retrieved.
32. Mednick S, Nakayama K, Stickgold R (2003). Sleep-dependent learning: a nap is as good as a night. Nature Neuroscience 6(7):697–698. DOI 10.1038/nn1078. PMID 12819785. Authors declared no competing financial interests; a grant identifier appears in the acknowledgements but the funding body was not retrieved. N not retrieved.