Is the 90-Minute Ultradian Rhythm Actually Supported by Evidence?
Written By Aftertone Team
Thursday, May 14, 2026
15 min read
Is the 90-Minute Ultradian Rhythm Actually Supported by Evidence?
The 90-minute focus cycle is real, but the evidence for its application to waking cognitive performance is weaker and more indirect than most productivity literature suggests. The underlying biological rhythm (the Basic Rest-Activity Cycle, or BRAC) was documented by Nathaniel Kleitman in the 1960s and later extended to waking states by Peretz Lavie and colleagues. But the research base for the specific 90-minute recommendation as an optimal focus block length draws on a mix of sleep research, waking physiology research, and practitioner observation that ranges from well-supported to loosely extrapolated. The honest summary: there is a real 90-minute biological rhythm, it likely influences cognitive performance, and 90-minute focus blocks are consistent with both the research and practitioner experience. The precise figure is not a settled scientific fact of the same status as the Zeigarnik effect or Gloria Mark's interruption findings.
Kleitman's Basic Rest-Activity Cycle
Nathaniel Kleitman, the sleep researcher who also co-discovered REM sleep, described the Basic Rest-Activity Cycle (BRAC) in the 1960s. The BRAC is a roughly 90-minute oscillation that Kleitman identified in sleep, in which the brain cycles between periods of active processing (REM) and periods of quieter maintenance (non-REM). Kleitman proposed that this rhythm continues during waking hours, producing alternating periods of higher and lower cognitive alertness and performance across the day.
Kleitman's waking BRAC hypothesis was based on preliminary observations rather than rigorous experimental evidence for waking cognitive performance. The sleep-based evidence for the 90-minute oscillation is solid. The extension to waking states as a deterministic 90-minute cycle in cognitive performance was more speculative in Kleitman's original formulation.
Lavie's research on waking ultradian rhythms
Peretz Lavie's research group at the Technion in Israel conducted more systematic studies of ultradian rhythms in waking performance across the 1980s and 1990s. Lavie identified what he called "sleepiness gates" and "wake maintenance zones": periods of relatively higher and lower sleep propensity within the waking day that had an approximate 90-to-120-minute periodicity. His research also documented corresponding fluctuations in cognitive performance and subjective alertness across similar periods.
Lavie's findings support the existence of an ultradian influence on waking alertness and performance, but the exact periodicity in his research was variable: 90 to 120 minutes, not a precise 90-minute cycle. Individual variation is substantial, and the amplitude of the oscillation (how much it affects performance) is less dramatic than the sleep cycle it mirrors. The practical implication is not a sharp 90-minute on/off switch but a gentler ebb and flow in cognitive performance that a well-designed work schedule can align with.
How the 90-minute recommendation entered productivity science
The specific 90-minute focus block recommendation was popularised primarily by performance researcher Peretz Lavie's work, K. Anders Ericsson's research on deliberate practice, and subsequent synthesis by writers including Tony Schwartz and Steven Kotler. Ericsson's research on expert performers found that elite performers in multiple fields (musicians, athletes, chess players) typically practised in sessions of approximately 90 minutes, with recovery periods between sessions. The sessions were not determined by the subjects' biology directly but by what they reported produced the best performance without fatigue accumulation.
The convergence of ultradian rhythm research, deliberate practice research, and expert practitioner observation around 90 minutes provides a coherent basis for the recommendation. But the research supporting it is observational, correlational, or extrapolated from related domains rather than produced by controlled experiments comparing 90-minute versus other focus block lengths on cognitive performance outcomes. The recommendation is well-grounded and consistent with available evidence; it is not derived from a clean experimental finding of the type that "90 minutes is optimal" would imply.
What the evidence actually supports
The evidence clearly supports that cognitive performance fluctuates across the day in a pattern with a period of roughly 90 to 120 minutes. The evidence supports that expert performers across domains tend to use approximately 90-minute focused work sessions with recovery between them. The evidence supports that sessions longer than 90 minutes tend to produce diminishing returns in performance quality and increasing fatigue.
The evidence is less clear on whether exactly 90 minutes is optimal for all people and all tasks, whether the ultradian rhythm directly determines the optimal session length or is simply correlated with what practitioners discover by experience, and whether the recovery periods between sessions need to be a specific length. The practical recommendation is: 90-minute sessions with breaks is a well-grounded default, individual calibration should refine it, and the principle (work in bounded sessions with recovery) is more firmly supported than the specific duration.
What the research does and does not establish
The research clearly establishes that a roughly 90-to-120-minute biological oscillation exists in both sleep and waking states (Kleitman; Lavie), that this rhythm corresponds to measurable fluctuations in alertness and cognitive performance across the waking day, and that expert performers across domains (Ericsson's deliberate practice research) tend to work in sessions of approximately 90 minutes with recovery between them.
The research does not establish that 90 minutes is the precise optimal focus block length for all people and all tasks. Lavie's waking research found a periodicity of 90 to 120 minutes — not an exact 90-minute cycle — with significant individual variation. The 90-minute recommendation is a convergence of related evidence streams rather than a single clean experimental finding of the form "90 minutes was compared to 60 and 120 and outperformed both." No such trial exists.
The research also does not establish the optimal length of recovery periods between sessions, the minimum block length required to benefit from the rhythm, or whether the cycle can be deliberately shifted. The biological rhythm is real; the specific practical recommendations derived from it carry more uncertainty than their confident citation in productivity writing suggests. A 90-minute default is well-grounded — but individual calibration from tracked experience is more reliable than assuming the population average applies.
Individual calibration is more reliable than the population average. Aftertone's planned versus actual tracking builds this calibration from your own data: comparing focus session length to output quality across several weeks reveals where your personal performance degrades — which may be at 75 minutes, or 110 minutes, rather than the population-average 90.
Frequently asked questions
Is the 90-minute ultradian rhythm a real finding?
The 90-minute ultradian rhythm is a real biological oscillation but its direct application to waking cognitive performance is less definitively established than its frequent citation suggests. Kleitman documented the Basic Rest-Activity Cycle as a roughly 90-minute oscillation in sleep. Lavie's research found corresponding waking alertness fluctuations with a periodicity of roughly 90 to 120 minutes. The rhythm is real; the specific 90-minute figure and its translation to optimal focus block length carry more uncertainty than most productivity writing acknowledges.
Should I use 90-minute focus blocks because of the ultradian rhythm?
90-minute focus blocks are a well-grounded recommendation consistent with ultradian rhythm research, deliberate practice research (Ericsson), and expert practitioner observation. They are not derived from a clean controlled experiment showing 90 minutes is optimal. Use 90 minutes as an evidence-consistent default, calibrate based on your own experience of when quality and focus begin to degrade, and treat the block and recovery structure as more important than the specific duration.
What is the Basic Rest-Activity Cycle?
The roughly 90-minute oscillation between active and quieter brain states that Nathaniel Kleitman identified in sleep research in the 1960s and proposed extends into waking states. In sleep, it corresponds to the REM/non-REM cycle. In waking states, Kleitman proposed it produces alternating periods of higher and lower alertness and performance. Lavie's subsequent research provided more systematic evidence for waking ultradian rhythms in alertness and cognitive performance.
How long should focus sessions actually be?
The research-consistent range is 60 to 90 minutes, with 90 minutes as a reasonable upper bound before performance quality typically begins to decline. The lower bound is determined by flow state entry time: deep work requires roughly 15 to 20 minutes to reach depth, which means blocks under 45 minutes rarely produce genuine deep work. Individual variation is significant; the optimal length for any person should be calibrated by tracking when focus quality degrades in practice rather than assumed from population averages.
Is the 90-minute focus recommendation as scientifically supported as other productivity findings?
Less so than the most-cited findings. Gloria Mark's 23-minute recovery figure, Gollwitzer's implementation intention meta-analysis, and the Zeigarnik effect all have direct experimental evidence. The 90-minute recommendation rests on a convergence of related evidence streams rather than a single clean experimental finding. It is well-grounded and consistent with available evidence, and the biological rhythm basis is real. The specific precision of "90 minutes" carries more uncertainty than its frequent citation suggests.
