Non-REM sleep
Sleep stages ranging from light sleep to slow-wave sleep. Brain activity, muscle tone and autonomic function change across these stages.
Insights / 07 · Sleep · 12 min read
How do pressure and timing shape sleep?
Sleep changes across the night, follows an internal clock and builds on time awake. It can influence attention, training quality and recovery, but no single number or wearable can describe it completely.
06 / Two processes
Pressure meets the body clock.
Sleep timing is influenced by accumulating sleep pressure and circadian timing. The lines show concepts, not an individual sleep test.
Structure / relationship
Illustrative process
Pressure
Homeostatic sleep pressure generally accumulates during wakefulness and dissipates during sleep. It is not simply a feeling of tiredness.
Clock
Circadian rhythms organise timing across roughly 24 hours. Light and routine help align that timing with the environment.
Architecture
Non-REM and REM stages recur during the night. A neat fixed-cycle graphic cannot describe every person or night.
Practice
Timing, opportunity, continuity and a workable routine matter. Persistent sleep difficulties need individual context rather than a generic score.
The short answer
Duration matters. Timing, regularity and quality matter too.
Sleep is not time switched off. The brain cycles through non-REM and REM sleep while the body regulates temperature, hormones, memory, immune activity and tissue repair. Most adults should regularly obtain at least seven hours, but need varies. A useful plan protects enough opportunity for sleep, keeps timing reasonably stable and looks at daytime function rather than chasing a perfect tracker score.
Two processes
Why sleepiness arrives.
Two interacting systems help organise sleep. Circadian timing is the roughly 24-hour rhythm coordinated by an internal clock in the brain. Light is its strongest time cue. The clock helps set when alertness, melatonin release and body temperature rise or fall.
Sleep pressure, sometimes called homeostatic sleep drive, generally builds while we are awake and falls during sleep. The longer a person stays awake, the stronger this pressure tends to become. Morning light, evening light, naps, caffeine, shift work and travel can alter how these systems line up. That is why being exhausted and being biologically ready to sleep are related, but not identical.
A simplified 24-hour model
Morning light anchors the clock
→
Sleep pressure rises while awake
→
Evening timing opens a sleep window
→
Sleep lowers pressure before the next day
Essential terms
REM sleep
A stage with rapid eye movements, vivid dreaming and a brain activity pattern distinct from non-REM sleep. REM periods tend to lengthen later in the night.
Sleep opportunity
The time reserved for sleep. Time in bed is not automatically the same as time asleep.
Sleep efficiency
The proportion of time in bed actually spent asleep. It can be estimated, but consumer devices have limits.
Chronotype
A person’s tendency towards earlier or later sleep timing, shaped by biology, age and environment.
Sleep debt
A practical label for accumulated insufficient sleep. It is not an exact bank balance that one long night always erases.
The night has architecture.
Sleep unfolds in repeated cycles rather than one uniform state. Non-REM sleep includes lighter stages and deeper slow-wave sleep. REM sleep has different brain and muscle patterns. A healthy night usually moves through these states several times. More slow-wave sleep tends to occur earlier; REM periods often become longer towards morning.
This matters because shortening the night at one end can change more than total minutes. It can also change which parts of the night are lost. Yet assigning one stage to one outcome is too simple. Memory, emotional regulation, metabolism and physical restoration emerge from interacting processes across the whole night.
Layer 1:
sleep has stages. Layer 2:
the distribution of stages changes across the night. Layer 3:
stage estimates from a wrist device are less precise than laboratory polysomnography, which measures brain activity, eye movement and muscle activity.
Study story · Van Dongen et al. · 2003
Small nightly losses accumulated.
48
healthy adults, 21–38 years
14
days of restricted sleep opportunity
4 / 6 / 8 h
time in bed per night
Researchers randomly assigned participants to 4, 6 or 8 hours in bed each night for two weeks; another group remained awake for up to three nights. Repeated tests measured sustained attention, working memory, processing speed and subjective sleepiness.
Performance deficits accumulated in a dose-dependent way. Six hours in bed did not behave like a harmless compromise across the full two weeks: attention lapses and other cognitive problems grew with repeated restriction. Participants’ own sleepiness ratings did not fully track the continuing performance loss.
What it means.
People may adapt to the feeling of short sleep faster than performance adapts. A demanding workday or a good training session can therefore feel manageable while attention is already less stable.
What it does not mean.
The sample was small, healthy and young-to-middle-aged, and time in bed is not identical to sleep obtained. The experiment does not prove that every person needs exactly eight hours or predict an individual athlete’s performance.
Sleep and physical performance.
Acute sleep loss can affect reaction time, perceived effort, attention and some measures of strength, speed or endurance. The size of the effect varies with the task, the amount and timing of lost sleep, time of day and the person studied. A 2022 systematic review and meta-analysis combined 69 publications and 227 performance outcomes. Overall performance was worse after acute sleep loss, especially when testing occurred later in the day, but studies were heterogeneous and participants were predominantly men.
That limitation matters. A meta-analysis combines studies to estimate an overall pattern; it does not make different protocols identical. Losing one entire night, shortening several nights and waking early can have different consequences. When methods and outcomes vary widely, the pooled result is a guide to direction and uncertainty rather than a personal forecast.
An endurance-specific review examined 31 studies with 478 participants. Sleep deprivation generally impaired endurance performance, with longer tasks appearing more vulnerable. Again, study quality, protocols and participant characteristics varied. The useful conclusion is cautious: insufficient sleep can reduce the capacity to perform and to judge effort, but there is no universal percentage penalty for one short night.
Study story · Mah et al. · 2011
More sleep opportunity in basketball players.
11
male varsity basketball players
5–7
weeks of sleep extension
10 h
target time in bed
After a two-to-four-week baseline, players were asked to spend at least ten hours in bed each night. Researchers tracked sleep, sprint time, shooting accuracy, reaction time, mood and daytime sleepiness.
During the extension period, measured sleep increased and several performance measures improved. The study is memorable because it asks a practical question: what happens when athletes with restricted schedules create more sleep opportunity?
It was also small, involved only men from one team and did not randomise players to a separate control group. Practice effects, the season and other changes could contribute. The result supports sleep opportunity as a performance variable; it does not establish ten hours as a universal prescription.
Recovery is broader than soreness.
Nervous system
Attention, reaction, learning and decision quality influence how well a skill is practised and how safely complex work is performed.
Metabolism
Sleep restriction can alter appetite, glucose regulation and hormonal signals. These responses depend on dose and context.
Tissue and immune function
Repair and immune processes continue around the clock. Sleep supports the conditions for recovery, but it does not replace nutrition or sensible training load.
Recovery means returning from the demands of training and life with the capacity to perform again. Sleep is one part of that process. A perfect sleep routine cannot compensate indefinitely for excessive training, inadequate energy intake, illness or psychological strain. Conversely, one disrupted night does not erase adaptation. Look at patterns.
A practical hierarchy
Protect the foundations before optimising details.
Priority
What to do
Why it comes first
1 · Opportunity
Reserve enough time for the sleep you need, including time to settle.
No supplement or tracker can create hours that are not available.
2 · Regularity
Keep wake time and sleep timing reasonably stable when life allows.
Regular timing supports circadian alignment.
3 · Light and stimulants
Seek daylight after waking; notice evening light and caffeine timing.
Light shifts the clock. Caffeine blocks adenosine receptors, can mask sleepiness and may disrupt later sleep.
4 · Environment
Create a dark, quiet and comfortably cool place.
Reduce avoidable interruptions rather than chasing perfect conditions.
5 · Feedback
Track daytime alertness, mood and training response alongside any device.
A wearable estimate is one data point, not a diagnosis.
Seven hours is a population threshold, not a stopwatch rule.
The American Academy of Sleep Medicine and Sleep Research Society recommend that adults regularly sleep seven or more hours for health. Athlete consensus guidance stresses that individual need and sport context vary.
When data helps—and when it distracts.
Consumer wearables estimate sleep from signals such as movement and heart rate. They may help reveal bedtime regularity or a broad trend. They do not directly measure sleep stages with the same method as polysomnography, and their algorithms differ. A large change can prompt a question; a small nightly fluctuation should not automatically prompt a decision.
If checking a score increases worry, treat the score as optional. Persistent difficulty falling asleep, repeated waking, loud snoring with pauses in breathing, severe daytime sleepiness or sleep problems lasting for months deserve professional assessment. Sudden changes can also relate to medication, illness, mood or work schedules.
From the Instagram archive · Evidence update · 4 October 2026
Caffeine has a dose.
Sleep has a context.
An afternoon coffee can improve alertness while making the next night less restorative. Caffeine binds to adenosine receptors: docking sites involved in sleep–wake regulation. Blocking a signal is different from removing the underlying need for sleep. Responses also differ between people and between occasional and habitual use. Reichert et al., 2022.
The original ZACH post showed timing curves from Gardiner and colleagues. Those curves are useful, but their cut-offs need careful interpretation.
Study lens 01 · A pooled model
24 studies. One estimated relationship.
The 2023 review found that caffeine reduced total sleep time by about 45 minutes on average across the included comparisons. Its model estimated an 8.8-hour interval for 107 mg of caffeine and 13.2 hours for 217.5 mg to avoid a predicted reduction in sleep duration. These are model outputs using example doses, not boundaries tested in every person. Coffee and supplement caffeine content also varies.
A cut-off is not a biological switch. The pooled average does not mean your coffee costs exactly 45 minutes, or that tea cannot affect sleep. Nor does a non-significant estimate establish zero effect. Gardiner et al. · Sleep Medicine Reviews · 2023.
Study lens 02 · A direct experiment
Same people. Different dose and timing.
23 men, mean age 25.3 years, with habitual caffeine intake below 300 mg/day completed a blinded, randomised crossover trial.
- 23 participants Each served as his own comparison.
- 7 conditions Placebo; 100 or 400 mg at 12, 8 or 4 hours before bed.
- 48-hour washout Conditions were separated to reduce carry-over.
- Sleep measured In-home partial polysomnography and sleep diaries.
100 mg:
No statistically significant objective or subjective sleep effect was detected at the tested times. That does not prove no individual effect.
See both results without interacting
100 mg: No significant effect detected. 400 mg: Sleep initiation and architecture were affected with intake within 12 hours; fragmentation increased within 8 hours. The small, male sample and single-dose design limit generalisation.
Use the evidence as a question.
Ask what you took, how much, when, and whether your sleep changed. If caffeine seems to interfere, an earlier or smaller intake is a reasonable variable to discuss or adjust. Keep sleep opportunity and timing in view too. A later small-dose trial and an earlier pooled timing model can differ without either being meaningless: they studied different doses, people and comparisons.
For the method behind that distinction, continue with How to read evidence. For the broader working week, explore Executive Health.
Keep these four ideas
Sleep has timing.
An internal clock and accumulated sleep pressure interact.
Sleep has structure.
Non-REM and REM sleep alternate across the night.
Restriction accumulates.
Repeated short nights can impair performance before the feeling seems severe.
Context decides.
Duration, regularity, daytime function, health and schedule belong in the same assessment.
Continue learning
Sources & references
Evidence used in this article.
01 · Van Dongen et al. · Sleep · 2003
Cumulative cost of additional wakefulness. Randomised laboratory restriction to 4, 6 or 8 hours in bed for 14 days. doi:10.1093/sleep/26.2.117
02 · Walsh et al. · British Journal of Sports Medicine · 2021
Expert consensus on sleep in athletes, including assessment and a practical sleep toolbox. doi:10.1136/bjsports-2020-102025
03 · Craven et al. · Sports Medicine · 2022
Systematic review and meta-analysis of acute sleep loss and physical performance. PubMed
04 · Lopes et al. · European Journal of Sport Science · 2023
Systematic review and meta-analysis of sleep deprivation and endurance performance across 31 studies. PubMed
05 · Mah et al. · Sleep · 2011
Sleep extension and athletic performance in collegiate basketball players. doi:10.5665/SLEEP.1132
06 · Watson et al. · Sleep · 2015
AASM and Sleep Research Society consensus recommendation for healthy adults. PubMed
07 · Reichert, Deboer & Landolt · Journal of Sleep Research · 2022
Adenosine, caffeine, and sleep–wake regulation: state of the science and perspectives. A narrative review of mechanisms, acute effects and habitual use. Read the review
ZACH / Connect the mechanism to the decision
What can you use this for?
01
Time awake
02
Circadian timing
03
Sleep opportunity
04
Daytime function
A reading sequence, not a measured causal model.
Keep the mechanism precise.
Sleep pressure generally builds during wakefulness, while circadian timing changes the biological tendency towards sleep and alertness across the day. Both can matter when exhaustion and the ability to fall asleep fail to line up.
A concrete way to apply the idea.
For a week, record bedtime, rising time, approximate sleep opportunity and how alert you feel during the day. Look for a scheduling constraint before chasing a stage score. Choose one feasible change, such as protecting a more consistent rising time, and review it against daytime function. Persistent sleep difficulties need a fuller assessment than a routine adjustment.
What we do not know.
Time in bed is not the same as sleep obtained. Wearable stage estimates do not have the same precision as laboratory polysomnography. A group result from restricted sleep cannot determine your exact need.
Continue learning / Follow the question
For a structured self-guided format, explore ZACH Learning. Check its current availability and scope before choosing a programme.