Ultradian Rhythm: Is the 90-Minute Focus Cycle Real?
Ultradian Rhythm: Is the 90-Minute Focus Cycle Real?
An ultradian rhythm is any biological cycle shorter than 24 hours: your heartbeat, hormone pulses, the roughly 90-minute rotation of sleep stages. That sleep-stage cycle is well measured. The claim that the same 90-minute rhythm controls your focus while awake is a different story. Some studies found oscillations in that range. More careful tests found none. Blocks of 60 to 90 minutes still make sense for deep work, but the reason is about how effort depletes, not a body clock.
Key Takeaways
- The 90-minute cycle is solid in sleep. You first enter REM sleep about 90 minutes after falling asleep, and the cycle repeats through the night.
- During waking hours, the evidence splits. One 1979 study found 60-to-100-minute oscillations. A 1995 study of 60 people tested every 10 minutes for 9 hours found no 90-minute periodicity at all.
- The "experts work in 90-minute sessions" story traces to a practice study of violinists, not a body-clock study. Sessions averaged 80 minutes, and the paper frames the limit as effort, not a timer.
- A 60-to-90-minute block is a reasonable starting point, but you should measure yours rather than assume it.
What Is an Ultradian Rhythm?
The word covers every recurring biological cycle with a period under 24 hours. A heartbeat, pulsatile cortisol secretion, and the cycling of NREM and REM sleep stages all qualify. Contrast that with circadian rhythms (about 24 hours, like your core body temperature curve) and infradian rhythms (longer than a day, like the menstrual cycle).
In sleep research, "ultradian" almost always means the 90-to-120-minute cycle of sleep stages. That cycle is the origin of the number productivity writers borrowed.
Where the 90-Minute Focus Claim Comes From
The idea that you should work in 90-minute blocks has been repeated so often that people assume it is settled science. It is not.
| Source | Year | What It Measured | Sample | What It Supports |
|---|---|---|---|---|
| NREM-REM sleep cycle | Ongoing | Sleep-stage cycling | Large body of research | ~90-min periodicity during sleep |
| Kleitman, BRAC hypothesis | ~1960, revisited 1982 | Hypothesis: sleep cycle extends into waking | Theoretical | The idea, not the proof |
| Klein & Armitage, Science | 1979 | Verbal and spatial tasks every 15 min, 8 hours | Not stated | 60-to-100-min oscillations, 180° out of phase |
| Kripke et al. | 1977 | Ambulatory monitoring symposium paper | Not stated | Titled "There's no basic rest-activity cycle" |
| Neubauer & Freudenthaler, Biol Psychol | 1995 | Sentence verification + alertness every 10 min, 9 hours | 60 people | No 90-min periodicity for any variable |
| Ericsson et al., Psych Review | 1993 | Violin students' practice structure | 30 (Study 1) | Sessions averaged 80 min; about effort, not a clock |
| Schwartz, New York Times | 2013 | Essay popularizing the concept | N/A | "Human beings operate in 90-to-120-minute cycles" |
| DeskTime app tracking | 2014, 2020, 2021 | Time-tracking data, top 10% of users | App users | 52/17, then 80/17, then 112/26; not peer reviewed |
Is There a 90-Minute Focus Cycle While You Are Awake?
Nathaniel Kleitman proposed in the early 1960s that the 90-minute cycle running through sleep continues during the day. He called it the basic rest-activity cycle (BRAC).
Some data pointed his direction. Klein and Armitage (1979) had subjects perform verbal and spatial tasks every 15 minutes for 8 hours and found 60-to-100-minute oscillations, with verbal and spatial scores 180 degrees out of phase.
Then the nulls came. Kripke and colleagues titled a 1977 paper "There's no basic rest-activity cycle." And Neubauer and Freudenthaler (1995) ran the most thorough test: 60 people tested every 10 minutes for 9 hours on a sentence-verification task, plus alertness, mood, and heart-period measurements. Spectral analysis found no 90-minute periodicity for any variable. Longer cycles were the main sources of variance. Their conclusion: several earlier positive reports "might be due to lack of conservatism in statistical methods."
Honest summary: possible, inconsistent, too weak to set your calendar by.
Where "Work in 90-Minute Blocks" Really Comes From
The phrase got its current life from two sources, neither of which is a body-clock study.
Ericsson's violinists. The 1993 deliberate-practice paper in Psychological Review studied 30 violin students at the Music Academy of West Berlin. Practice sessions averaged 80 minutes with no reliable difference among skill groups. A second study found expert pianists practiced in sessions of 87.9 minutes versus 46.0 for amateurs. The paper's real finding: training studies showed "essentially no benefit from durations exceeding 4 hr per day," and the violinists napped to recover, the best groups more than the music-education students. Session length tracked the depletion of concentration, not a biological timer.
Schwartz's popularization. Tony Schwartz's 2013 New York Times essay translated Ericsson's data and Kleitman's hypothesis into one actionable claim: work in 90-to-120-minute cycles, then rest. DeskTime's app data added numbers: their top 10% most productive users worked 52 minutes / broke 17 minutes in 2014, shifted to 80/17 in early 2020, then 112/26 in 2021. The "ideal" block moved with how people worked. None of it was peer reviewed.
Effort Runs Down and Breaks Help
You do not need a 90-minute clock to justify breaks. Effort itself is the limiting factor.
A 2022 meta-analysis in PLOS ONE (Albulescu et al.) found micro-breaks produced small gains in vigor (d = 0.36) and reduced fatigue (d = 0.35). The overall effect on performance was not statistically significant (d = 0.16, p = .116), reaching significance only for less demanding tasks. Longer breaks gave bigger performance boosts, and recovering from depleting tasks "may need more than 10-minute breaks."
A real daily rhythm does exist, on a circadian scale: Lavie (1986) documented a secondary mid-afternoon sleepiness peak. If your focus craters after lunch, circadian biology is a likelier culprit than an ultradian cycle. (See our guide to circadian rhythm, chronotypes and the post-lunch dip or how to keep doing deep work after lunch.)
How Long Should a Focus Block Be? A One-Week Test
If there is no reliable 90-minute clock, you need to find your own number.
Start at 60 Minutes
For demanding cognitive work, set a timer for 60 minutes. If quality holds when it rings, extend by 15-minute increments next session. For lighter work, start at 90 and adjust down.
Log Three Things Per Block
Write down the start time, the minute you first drifted, and your output (pages, tickets, problems solved). After 5 to 10 blocks, your median drift time is your personal block length.
Scale Breaks to Effort
After a moderate block, 5 minutes works. After 90 minutes of the hardest work you do, take 15 to 20. Movement and daylight both help.
Cap Deep Blocks at 3 to 4 Per Day
Ericsson's review found essentially no benefit from deliberate practice beyond about 4 hours a day. Put the hardest block during your best hours. Protect it from interruptions, because context switching has a real cost (see what context switching really costs). For a full daily layout, try a time-blocking template built around energy lanes.
Ultradian Rhythm and Focus: The Bottom Line
The 90-minute number is a sleep fact that got stretched into a productivity rule. The waking evidence is thin and contradicted by the most careful study on record. What is real: effort depletes, breaks help recovery, and the useful length of a focused work block varies by person and task. Measure your block length. Do not assume it.
Frequently Asked Questions
What is an ultradian rhythm in simple terms?
An ultradian rhythm is any biological cycle that repeats in less than 24 hours. Your heartbeat, hormone pulses, and the cycling of sleep stages (roughly every 90 to 120 minutes) are all ultradian rhythms. The term shows up most in sleep science, where the NREM-REM cycle is the classic example.
What is the difference between ultradian and circadian rhythms?
Ultradian rhythms cycle faster than once per day. Circadian rhythms run on an approximately 24-hour period, like your sleep-wake cycle and core body temperature curve. A third category, infradian, covers rhythms longer than a day.
Is the 90-minute focus cycle scientifically proven?
Not convincingly. Klein and Armitage (1979) found 60-to-100-minute oscillations in cognitive tasks, but Neubauer and Freudenthaler (1995) tested 60 people every 10 minutes for 9 hours and found no 90-minute periodicity. The most careful study is a null result.
How long can you actually focus on one task?
It depends on the task and the person. Ericsson's data showed violin students practicing in sessions averaging 80 minutes, and the training studies he reviewed found little benefit beyond about 4 hours of effortful practice a day. Your personal limit is best found by logging when you first drift during timed blocks.
How long should a break be after a 90-minute work block?
Albulescu et al. (2022) found longer breaks produced bigger performance gains and that recovering from highly depleting tasks may need more than 10-minute breaks. A 15-to-20-minute break after a hard 90-minute block is a reasonable starting point.
What is the basic rest-activity cycle (BRAC)?
The BRAC is a hypothesis proposed by Nathaniel Kleitman in the early 1960s. He suggested the approximately 90-minute cycle of sleep stages continues during waking hours, producing alternating phases of higher and lower alertness. Some studies found supporting oscillations; others found none. It remains a hypothesis.
One Block, One Pouch, Measured Your Way
Nobody can hand you a universal focus timer. You run the blocks and find where your concentration actually breaks. The tool question is simpler: what helps you start sharp and stay there?
Caffeine plus L-theanine is a well-studied pairing for attention. A 2010 trial (n = 44) found 97 mg L-theanine and 40 mg caffeine improved task-switching accuracy and self-reported alertness versus placebo (Giesbrecht et al.; Unilever-funded, dose differs from Roon's). A 2025 review of 50 RCTs found the combination favored attention-switching accuracy at hour 2 (SMD 0.33; Lipton/Unilever-linked authors).
Roon puts 80 mg caffeine, 60 mg L-theanine, 25 mg methylliberine (Dynamine), and 5 mg theacrine (TeaCrine) in a zero-nicotine pouch. Onset runs 5 to 10 minutes, so the pouch doubles as a start cue. In an n = 12 crossover PK study, 100 mg methylliberine with 150 mg caffeine cut caffeine's oral clearance from 41.9 to 17.1 L/hr and extended its half-life from 7.2 to 15 hours, with no increase in peak concentration. That is 4x Roon's methylliberine and roughly 1.9x its caffeine, so the direction applies but the magnitudes do not transfer. Roon's felt window is 6 to 8 hours, enough for several blocks. Stop at least 8 hours before bed.
Roon is not a substitute for sleep, breaks, or block structure. From $9.16 per tin on a 12-tin subscription. Try it for your next block.
Related from Roon
- The Neuroscience of Flow
- What Context Switching Really Costs
- Pomodoro and Where Caffeine Fits
- What a Short Nap Does for Alertness
Written by Roon Team
