Ultradian rhythm: what the 90 minute cycle really is, and what the research supports

An ultradian rhythm is any biological cycle that repeats more than once a day. The best known is the roughly 90 to 100 minute cycle of sleep stages at night. Productivity writing has turned that into a simple rule: your brain runs in 90 minute waves all day, so work for 90 minutes and then rest. The real research is more interesting and much less tidy. Some lab studies found daytime cycles of about 100 to 120 minutes, others found nothing at the 90 minute mark, and the period changes with the measure, the person and the method used to find it.

This guide explains what an ultradian rhythm is, where the 90 minute idea came from, what the daytime studies actually found, where the popular work rule goes beyond the evidence, and how to find a work and break pattern that fits you rather than a number from a blog post.

Short answer: ultradian rhythms are real, and the night time sleep cycle is well established. A fixed 90 minute focus cycle while you are awake is a hypothesis with mixed support, not a proven law. Treat 60 to 120 minute blocks with short breaks as a reasonable starting point to test, not a biological rule.

What is an ultradian rhythm?

Biologists sort body rhythms by how long one cycle takes. A circadian rhythm takes about a day. An ultradian rhythm is shorter than a day, so it repeats several times within 24 hours. An infradian rhythm is longer than a day.

Rhythm type	Cycle length	Common examples
Ultradian	Shorter than 24 hours (minutes to hours)	Night time sleep stage cycles, pulses of hormone release, heartbeat and breathing in the broadest sense
Circadian	About 24 hours	The sleep and wake cycle, core body temperature, the early afternoon dip in alertness
Infradian	Longer than 24 hours	Monthly and seasonal cycles

Ultradian rhythms are not fringe science. Hormone systems are a clear example. A review by Lightman and Conway-Campbell in Nature Reviews Neuroscience describes how stress hormone output is not a steady level but a series of pulses, and argues that this constant oscillation helps the brain and body stay responsive (Lightman and Conway-Campbell, 2010). So the question is not whether the body has cycles shorter than a day. It does. The question is whether there is a reliable focus cycle of about 90 minutes while you are awake, and whether you can plan your work around it.

Where the 90 minute idea comes from

The idea starts with sleep. During the night you move through repeating cycles of non-REM and REM sleep. In one lab study, Lavie and Zomer found that the dominant frequency of those night time cycles was about 14.4 cycles per day, which works out to roughly 100 minutes per cycle (Lavie and Zomer, 1984). That is where the familiar "90 minute sleep cycle" comes from, although real cycles vary from person to person and across the night.

Nathaniel Kleitman, one of the founders of sleep research, went a step further. Around 1960 he proposed a basic rest-activity cycle: the same ultradian rhythm that drives sleep stages at night might keep running during the day as a rise and fall in activity and alertness. His 1982 review, titled "Basic rest-activity cycle, 22 years later", gathered data in support of the idea that this cycle shows up both in the alternation of sleep stages and as a similar periodicity in wakefulness (Kleitman, 1982).

That is an important point that most popular articles skip. Kleitman described a hypothesis and the evidence he thought supported it. He did not test a 90 minute work schedule, and he did not show that working in 90 minute blocks makes anyone more productive. The work rule came later, from productivity writing that combined his idea with other findings, such as the practice sessions of expert musicians discussed in our guide to deep work.

What daytime studies actually found

Researchers have tried to find Kleitman's daytime cycle in several ways: by measuring how easily people fall asleep, how well they perform simple tasks, how alert they feel, and what their EEG shows. The results point in different directions.

Evidence for a daytime cycle

Peretz Lavie and colleagues used an "ultrashort" schedule in the sleep lab. Volunteers spent 15 minutes awake and then 5 minutes trying to sleep, over and over. In one study of nine young adults kept on this schedule for 12 hours, the amount of light (stage 1) sleep in each attempt rose and fell at about 14.4 cycles per day, roughly every 100 minutes. After a night of REM sleep deprivation or total sleep deprivation, that 100 minute pattern was disrupted. Deeper stage 2 sleep did not follow an ultradian pattern; it clustered in the mid-afternoon (Lavie and Scherson, 1981).

A Japanese study measured ten male university students every 15 minutes for nine hours, recording task performance, mood, self-rated performance and EEG. Task performance and the subjective ratings moved together at about 12 cycles per day, roughly every two hours. EEG activity followed different patterns, a slower cycle of about 6 per day and faster ones of 10 to 18 per day. The authors read this as support for several ultradian rhythms running at once rather than one master clock (Hayashi, Sato and Hori, 1994).

Evidence that complicates it

In a second ultrashort schedule study, the main rhythm in sleepiness across the afternoon, evening and morning ran at about 7.2 cycles per day, roughly every 200 minutes, not 100. A 100 minute peak only appeared when each morning's data were lined up with the person's last REM period of the night (Lavie and Zomer, 1984). In other words, a short cycle could be found, but only under specific conditions and analysis choices.

Evidence against a 90 minute performance cycle

The largest single study in this set tested 60 people every 10 minutes for nine hours on a simple sentence checking task, with ratings of alertness and mood and a heart measure taken during each test. Using conservative statistics, the researchers found no significant 90 minute rhythm in any of the measures. Longer cycles explained more of the variation. When they reanalysed the data in other ways, a 90 minute cycle could sometimes appear, which led them to suggest that some earlier positive reports may reflect statistical methods that were not strict enough (Neubauer and Freudenthaler, 1995).

Study	Who and how	Cycle found	Main limitation
Lavie and Scherson, 1981	9 young adults, 15 minutes awake and 5 minutes sleep attempt for 12 hours	About 100 minutes in light sleep	Small sample; measures sleepiness in a lab, not work output
Lavie and Zomer, 1984	8 adults on the same kind of schedule	About 200 minutes; 100 minutes only after aligning to the last night time REM period	Small sample; result depends on alignment
Hayashi, Sato and Hori, 1994	10 male students, measured every 15 minutes for 9 hours	About 2 hours in performance and mood; different cycles in EEG	Very small, all male sample
Neubauer and Freudenthaler, 1995	60 people, tested every 10 minutes for 9 hours	No significant 90 minute cycle	One simple task; lab setting

Why the cycle length keeps changing

Across these studies the "ultradian cycle" comes out at about 100 minutes, about 120 minutes, about 200 minutes, or not at all. There are a few likely reasons.

• Different measures follow different rhythms. Hayashi and colleagues found one cycle in performance and mood, and other cycles in EEG. A single "focus wave" is too simple a picture.
• Method matters. Finding a rhythm in noisy data means choosing how to smooth it, detrend it and test it. Neubauer and Freudenthaler showed that a 90 minute cycle could appear or disappear depending on those choices.
• Small samples. Most positive studies tested eight to ten people. Individual cycles may be real but vary too much between people to average into one number.
• Daily life is not a lab. These studies controlled meals, light, posture and tasks. At work, meetings, messages, food, caffeine and the task itself push alertness around far more than a faint cycle might.

The bigger pattern is daily, not every 90 minutes

When researchers looked at sleepiness across a full 24 hours, the clearest structure was circadian. In three experiments using the ultrashort schedule after a night without sleep, sleepiness had a large night time peak and a smaller peak in the mid-afternoon, with a "forbidden zone" of low sleep pressure in the evening, around 8 to 10 pm (Lavie, 1986). That mid-afternoon peak is the dip many people feel after lunch, which we cover in our guide to the afternoon slump.

For planning your day, that matters more than any 90 minute count. Your hardest thinking is usually easier to protect in the hours when you are reliably alert, and harder to force during your daily low point, whatever the exact minute marks inside those hours look like.

Where the popular 90 minute work rule overreaches

Many articles on ultradian rhythms say some version of the following. Here is how each claim holds up against the studies above.

• "Kleitman discovered that we work in 90 minute cycles." He proposed that a sleep-related cycle might continue during the day and gathered evidence for it. He did not show a 90 minute work cycle.
• "Every 90 minutes your brain needs a 20 minute rest." None of the studies here tested a rest length. The 20 minute figure is a convention, not a finding.
• "You will feel a 45 minute high followed by a 45 minute low." No study here reported a fixed half and half split like this.
• "Working with your ultradian rhythm boosts productivity." No study here tested work output under a 90 minute schedule against another schedule.
• "Everyone's cycle is 90 minutes." Positive studies found about 100 to 120 minutes, one found 200, and the largest single study found no 90 minute cycle at all.

None of this means 90 minute blocks are a bad idea. It means the reason to use them is practical, not that biology demands exactly 90 minutes.

What the research on breaks does support

If the exact cycle is uncertain, the value of breaks is a more useful question. A 2022 meta-analysis pooled 22 study samples (2,335 people) on micro-breaks of up to 10 minutes. Micro-breaks gave small but reliable increases in vigour and decreases in fatigue. Their effect on overall performance was small and not statistically significant, with benefits showing up mainly on less demanding tasks. Longer breaks were linked to a bigger performance boost, and the authors suggest that recovering from highly draining work may need more than 10 minutes (Albulescu and colleagues, 2022; full text).

So breaks help people feel less tired and more energetic, which matters on its own. For hard thinking, a two minute pause may not be enough, and a longer real break after a long block may be worth more. Our guide to brain breaks for adults has ideas for what to do with that time.

How to use ultradian rhythm thinking without the myth

1. Start with a block length, not a biology claim. Try 60 to 90 minutes of single-task work. If you often fade earlier, start at 45. The aim is a length you can protect, not a magic number.
2. Put your hardest work in your most alert hours. For many people that is late morning. Avoid scheduling demanding work in your usual afternoon dip if you can.
3. Watch for your own fade signals. Rereading the same paragraph, reaching for your phone, small errors and restlessness are practical signs a block has run its course. Stopping a few minutes early at a clean point is better than pushing through on autopilot.
4. Take a real break after a long block. Stand up, move, get water or step outside. Scrolling email is a task switch, not a break, and leaves attention residue.
5. Protect the block from interruptions. A rhythm does not help if notifications break it every ten minutes. Batch messages between blocks, as described in our guide to task batching.
6. Plan the blocks on your calendar. Time blocking gives each long session a place in the day so it is not squeezed out.

A two-week test to find your own work rhythm

Because published cycle lengths vary so much, your own record is the most useful evidence you can get.

• Days 1 to 5: work in your usual way, but note the start time of each focused session, when you first notice fading, and when you stop.
• Days 6 to 10: set a fixed block length (for example 75 minutes) with a 10 to 15 minute break away from screens between blocks. Note fade times again.
• Days 11 to 14: adjust the block length toward the point where you usually started fading, and keep the breaks.
• Review: compare finished work, error rates and how tired you felt at the end of each day. Look for a typical fade time and the hours when blocks went best. If there is no clear pattern, that is a real result too: your day may be shaped more by sleep, meetings and task type than by a fixed cycle.

Ultradian rhythm vs other work methods

Method	Typical block	What it is based on	Best for
Ultradian style blocks	60 to 120 minutes, then a longer break	Kleitman's hypothesis plus practical habit; mixed evidence for the cycle itself	Long, demanding tasks that need warm up time
Pomodoro	25 minutes, then 5 minutes	A timer convention, not a biological rhythm	Starting tasks you are avoiding; tasks that split easily
Time blocking	Any length, planned in advance	Planning and protecting time on a calendar	Busy days with meetings and competing tasks
Deep work	Often 60 to 90 minutes, a few hours a day at most	Rules of thumb drawn from expert practice studies	Skill building and hard, high value work

These approaches are not rivals. Many people use a Pomodoro timer to get started, then let a session run into a longer block once they are absorbed.

Where Zone fits

Zone is a pair of EEG earbuds with an app that shows app brain states, such as Locked In, In Flow, Drifting, Running Hot and Depleted. The earbuds read signals through an in-ear EEG array, described in our explainer on in-ear EEG. If you run the two-week test while wearing Zone, you can set session summaries beside your own notes and see whether shifts toward Drifting or Depleted tend to arrive at a similar point in your work blocks, or at similar times of day. Our guides to focus tracking and how to measure focus explain more. Treat these as app estimates, not a measurement of an ultradian rhythm or proof of your ideal block length, and keep your own notes and finished work as the main evidence. Zone does not diagnose, treat, or manage any medical condition.

Common questions

What is an ultradian rhythm in simple terms?

A biological cycle that repeats more than once in 24 hours. Night time sleep stage cycles and the pulsed release of some hormones are common examples.

How long is an ultradian rhythm?

It depends on the rhythm. Night time sleep cycles run at roughly 90 to 100 minutes. Daytime studies have reported cycles of about 100 minutes, about 2 hours and about 200 minutes, and the largest study in this guide found no 90 minute cycle in performance at all.

What is the difference between ultradian and circadian rhythms?

A circadian rhythm takes about 24 hours, like the sleep and wake cycle. An ultradian rhythm is shorter and repeats several times a day. For daytime alertness, the circadian pattern, including the mid-afternoon dip, is the better established of the two.

Is the 90 minute work cycle real?

Partly. Kleitman's idea of a daytime rest-activity cycle has some lab support, but the cycle length varies between studies and one larger study found no 90 minute rhythm in performance. Working in 60 to 90 minute blocks can still be a useful habit; it just is not a proven biological law.

How do I find my own ultradian rhythm?

No simple test exists. A practical approach is to log when you start focused work, when you first notice fading and when you stop, over one to two weeks, then look for a typical fade time and your best hours of the day.

Sources

1. Kleitman, N. (1982). Basic rest-activity cycle, 22 years later. Sleep, 5(4), 311 to 317. doi.org/10.1093/sleep/5.4.311
2. Lavie, P., and Scherson, A. (1981). Ultrashort sleep-waking schedule. I. Evidence of ultradian rhythmicity in "sleepability". Electroencephalography and Clinical Neurophysiology, 52(2), 163 to 174. PubMed record
3. Lavie, P., and Zomer, J. (1984). Ultrashort sleep-waking schedule. II. Relationship between ultradian rhythms in sleepability and the REM-non-REM cycles and effects of the circadian phase. Electroencephalography and Clinical Neurophysiology, 57(1), 35 to 42. PubMed record
4. Lavie, P. (1986). Ultrashort sleep-waking schedule. III. 'Gates' and 'forbidden zones' for sleep. Electroencephalography and Clinical Neurophysiology, 63(5), 414 to 425. PubMed record
5. Hayashi, M., Sato, K., and Hori, T. (1994). Ultradian rhythms in task performance, self-evaluation, and EEG activity. Perceptual and Motor Skills, 79(2), 791 to 800. PubMed record
6. Neubauer, A. C., and Freudenthaler, H. H. (1995). Ultradian rhythms in cognitive performance: no evidence for a 1.5-h rhythm. Biological Psychology, 40(3), 281 to 298. PubMed record
7. Lightman, S. L., and Conway-Campbell, B. L. (2010). The crucial role of pulsatile activity of the HPA axis for continuous dynamic equilibration. Nature Reviews Neuroscience, 11(10), 710 to 718. PubMed record
8. Albulescu, P., Macsinga, I., Rusu, A., Sulea, C., Bodnaru, A., and Tulbure, B. T. (2022). "Give me a break!" A systematic review and meta-analysis on the efficacy of micro-breaks for increasing well-being and performance. PLOS ONE, 17(8), e0272460. PubMed record, full text