What Time Should You Go to Bed? The 90-Minute Rule, Checked

Written by the Nuvirox Research Team

Short answer: pick a wake time, count back seven to nine hours, add about 20 minutes for falling asleep — and then stop optimising. The popular sleep calculators are built on a tidy 90-minute cycle that real sleep does not reliably produce. The underlying idea, that waking mid-deep-sleep feels awful, is sound. The arithmetic built on top of it is more precise than the biology justifies, and the variable with the strongest evidence behind it is not what time you go to bed at all. It is whether you do the same thing tomorrow.

Where does the 90-minute rule come from?

From genuine research. The cyclical structure of sleep — a repeating progression through light non-REM, deep non-REM and REM — was established in the 1950s and 60s, and the average interval between REM periods in healthy adults lands somewhere close to 90 minutes. That is a real finding.

The problem is what happened next. An average across thousands of recordings became, in popular usage, a fixed unit of time that every person's every night obeys. Reported cycle durations range from roughly 70 to 120 minutes, they vary between people, and — the detail that breaks the calculators — they vary within a single night. Early cycles tend to be shorter and heavy with deep sleep; later ones stretch as REM periods expand. If your first cycle runs 75 minutes and your fourth runs 105, a calculator assuming five identical 90-minute blocks has misplaced your fifth boundary by nearly an hour.

Illustrative representation of the reported range. Because cycles also lengthen across a single night, a calculator assuming five identical 90-minute blocks will drift by the fourth one.

Does waking between cycles actually help?

The underlying phenomenon is real and is called sleep inertia: the disorientation and impaired performance that follows waking, which is markedly worse when you are pulled out of slow-wave sleep than out of light sleep. Anyone who has been woken from deep sleep by a phone call knows this from the inside.

So the principle holds. The implementation does not. To wake at a light-sleep boundary you need to know where the boundary is, and predicting that in advance from a fixed formula is not possible with the precision the calculators imply. Devices that offer a smart-alarm window attempt to detect it in real time instead, which is a more sensible approach in principle — but consumer devices infer stages from movement and heart rate rather than brain activity, and their stage detection has meaningful error, a limitation we went through in what your sleep tracker actually knows.

There is also a trade-off nobody mentions. Deliberately cutting a night short to land on a notional cycle boundary means sacrificing real sleep to avoid temporary grogginess. Sleep inertia dissipates within roughly 15 to 30 minutes. The lost hour does not.

What human studies actually show

Regularity beat duration in the largest objective study of the question. Windred and colleagues calculated a Sleep Regularity Index from more than 10 million hours of accelerometer data in 60,977 UK Biobank participants, with mortality followed for up to 7.8 years. Across the top four regularity quintiles compared with the least regular quintile, higher regularity was associated with 20% to 48% lower all-cause mortality risk, with similar patterns for cancer and cardiometabolic mortality. Critically, regularity was a stronger predictor of all-cause mortality than duration was.

Duration still matters, in a U-shape. A meta-analysis of prospective studies found both short and long habitual sleep associated with higher all-cause mortality, with the lowest risk clustering around seven hours. So the seven-to-nine-hour target is not arbitrary — it is just less discriminating than the regularity finding.

The honest counterweight: this is observational data. Both studies are cohorts, not experiments. Irregular sleep is entangled with shift work, illness, caregiving, poverty and mood disorders, any of which independently affect mortality. Nobody has randomised people to regular versus irregular schedules for a decade, and nobody will. The finding is strong enough to act on and not strong enough to call causal.

What a bedtime calculation won't fix

It will not create sleepiness at the hour you have selected. If your circadian phase runs late, going to bed at 10 p.m. because a calculator said so produces an hour of lying awake, which over enough repetitions builds the conditioned arousal that turns a timing problem into an insomnia problem.

It will not help if the real issue is fragmentation rather than scheduling. Untreated sleep-disordered breathing, restless legs, nocturia, pain and reflux all break sleep in ways no bedtime arithmetic touches. If you are in bed the right number of hours and still wake unrefreshed most days, that pattern deserves a clinical look rather than a better formula.

A version that actually works

Start from the wake time, because that is the anchor you cannot move on weekdays anyway. Count back the amount of sleep you need — most adults land between seven and nine hours, and the way to find yours is a stretch of unrestricted days, not a calculator. Add roughly 15 to 25 minutes for falling asleep. That is your target bedtime, and it is a target, not a deadline: going to bed when you are not yet sleepy is worse than going to bed 20 minutes late.

Then hold the wake time steady, including at weekends. Drifting two hours later on a Saturday produces the same phase shift as flying west, which is exactly the mechanism behind social jet lag. Get bright light shortly after waking to reinforce the anchor — the evidence for that is summarised in morning sunlight and sleep. And if the schedule keeps failing because you cannot fall asleep at the hour you have chosen, the answer is usually to move bedtime later temporarily, not earlier.

Frequently asked questions

Is 6 hours better than 7.5 if it lands on a cycle boundary? No. Total sleep does far more work than boundary timing, and cutting 90 minutes to avoid 20 minutes of grogginess is a poor trade.

Are sleep calculator apps useless then? Not useless — they are a reasonable prompt to work backwards from your wake time, which most people never do. Treat the output as approximate rather than as a set of five precise options.

What if my work schedule makes regularity impossible? Then aim for regularity within the pattern rather than across it, and accept a genuine constraint rather than a personal failing. Shift workers face a structural problem, not a discipline problem.

Does napping break the calculation? Naps reduce sleep pressure, so a long or late nap will push your sleep onset later. Short early-afternoon naps are much less disruptive than long ones after about 3 p.m.

From Nuvirox

Why we formulated Sleep+ Restore.

Most sleep formulas pick one lever. Sleep+ Restore was built around the fact that the research points at several at once: timing signals, the amino-acid precursors your brain uses to build them, and the calming botanicals that have actual human trial data behind them.

• 10 mg melatonin — a timing signal, not a sedative. Worth saying plainly: published trials generally use 0.3–5 mg, and meta-analysis puts melatonin's average effect on falling asleep at roughly seven minutes. If you are melatonin-sensitive, start with a lower-dose product.
• 905 mg Sleep Formula blend — L-tryptophan, L-theanine, chamomile, lemon balm, passionflower, hops, ashwagandha, Chinese skullcap, goji, GABA, taurine, inositol, St. John's Wort and 5-HTP, the botanicals most often studied in human sleep and relaxation research.
• Vitamin B6, calcium and magnesium — cofactors in the tryptophan-to-serotonin-to-melatonin pathway, included at nutritional rather than pharmacological amounts.
• 60-day money-back guarantee — long enough to actually evaluate it the way the research says you should, across weeks rather than a single hopeful night.

Two capsules, 30 servings per container. St. John's Wort and 5-HTP interact with a long list of prescription medicines — antidepressants, hormonal contraceptives, immunosuppressants, anticoagulants and more. Check with a pharmacist before starting if you take anything regularly.

Learn more about Sleep+ Restore →

The bottom line

The 90-minute rule takes a real average and treats it as a fixed law, which it is not. Waking out of deep sleep genuinely feels worse, but you cannot schedule around a boundary you cannot locate, and shortening your night to try is a bad trade. The evidence points at something less satisfying and more useful: choose a wake time you can actually keep every day, give yourself enough hours ahead of it, and let bedtime settle where your sleepiness puts it.

References

1. Windred DP, Burns AC, Lane JM, Saxena R, Rutter MK, Cain SW, Phillips AJK. Sleep regularity is a stronger predictor of mortality risk than sleep duration: a prospective cohort study. Sleep. 2024;47(1):zsad253. PMID: 37738616. doi:10.1093/sleep/zsad253
2. Ohayon MM, Carskadon MA, Guilleminault C, Vitiello MV. Meta-analysis of quantitative sleep parameters from childhood to old age in healthy individuals: developing normative sleep values across the human lifespan. Sleep. 2004;27(7):1255–1273. PMID: 15586779. doi:10.1093/sleep/27.7.1255
3. American Academy of Sleep Medicine. The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications, Version 3. Darien, IL: AASM; 2023.
4. 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. doi:10.1093/sleep/33.5.585
5. Trauer JM, Qian MY, Doyle JS, Rajaratnam SMW, Cunnington D. Cognitive behavioral therapy for chronic insomnia: a systematic review and meta-analysis. Annals of Internal Medicine. 2015;163(3):191–204. PMID: 26054060. doi:10.7326/M14-2841

*These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease. This article is for informational purposes only and is not a substitute for professional medical advice.