You run through four to six sleep cycles of about 90 minutes a night. Each cycle contains light sleep (N1 and N2), deep sleep (N3) and dream sleep (REM). In healthy adults, roughly half the night is light sleep, 13 to 23 percent is deep sleep and 20 to 25 percent is REM.
You rarely find those percentages on Dutch sleep sites. The biggest NL source on the sleep cycle, the Hersenstichting, names the stages but gives no breakdown.
Hence this table.
Which sleep stages are there?
There are four stages: falling asleep (N1), light sleep (N2), deep sleep (N3) and dream sleep (REM). You do not move neatly from light to deep, but run through them in cycles that repeat all night. The ratio between the stages shifts as the night goes on.
| Stage | Share of the night | What happens |
|---|---|---|
| N1, falling asleep | About 5% | Transition between waking and sleeping, lasts a few minutes |
| N2, light sleep | About 45 to 55% | The bulk of your night, heart rate and temperature drop |
| N3, deep sleep | About 13 to 23% | Growth hormone release peaks, mostly in the first half of the night |
| REM, dream sleep | About 20 to 25% | Memory processing, increases towards morning |
Note how it is spread across the night. Your deep sleep sits mostly in the first two cycles, your REM mostly in the last two.
That has a practical consequence. Go to bed two hours later but wake at your fixed time, and you mainly cut dream sleep. Get up two hours earlier and the same happens.
How long is a sleep cycle?
About 90 minutes, with plenty of variation between people and between nights. Sleep six hours and you get roughly four cycles. Sleep eight and you get five to six. So every cycle you miss is not an abstract percentage but a block of recovery that never ran.
The Hersenstichting uses the same 90 minutes. What strikes me as a coach: that number gets abused for advice like "sleep in multiples of 90 minutes". Your biology is not that tidy.
Your cycles are not a train timetable. They are an average.
How does your sleep architecture change with age?
Sharply, and earlier than most people expect. In a JAMA study of 149 healthy men, the share of deep sleep fell from 18.9 percent in 16 to 25-year-olds to 3.4 percent in 36 to 50-year-olds. Dream sleep stayed fairly stable across that span and only declined after age fifty, by roughly 10 minutes per decade (PMID 10938176).
Picture two athletes of 24 and 44 who both spend eight hours in bed. On paper they sleep the same. In deep sleep the gap is roughly fivefold.
The same study saw evening cortisol rise after age fifty by about 19.3 nmol/l per decade. Sleep and cortisol move together, in both directions.
What does your tracker actually measure?
Movement, heart rate, heart rate variability and skin temperature. An algorithm infers a stage from that. Your watch does not look at your brain waves, while lab staging is based on exactly those. So the pie chart is an estimate, not a measurement.
In a comparison of seven consumer trackers against polysomnography, the devices were good at telling sleep from wake, with a sensitivity of 0.93 or higher. They were weak at recognising lying awake (specificity 0.18 to 0.54) and inconsistent at classifying stages, especially during disrupted sleep (PMID 33378539).
Translated to your morning: your watch knows fairly well that you slept, too readily assumes you slept while you lay awake, and guesses at which stage you were in.
I still use those numbers myself. I just look at the trend across two weeks, not at last night.
So what should you look at?
At your bedtime, your wake time and how you feel in your first heavy set. Those are the data you can steer. The stage percentages are interesting, but there is literally nothing you can do about them.
If you want hard data next to your watch, blood is the addition your tracker is missing. Cortisol, glucose and testosterone show what is happening in the systems around your sleep.
The honest comparison is in sleep tracker versus blood values. How you protect your deep sleep is in deep sleep. The complete plan is in better sleep for athletes.
Our Whoop blood test was built to put exactly those two sides side by side.
My advice: look at your stage percentages once this week, then leave them alone until next month.
References
- Van Cauter E, Leproult R, Plat L. Age-related changes in slow wave sleep and REM sleep and relationship with growth hormone and cortisol levels in healthy men. JAMA. 2000. PMID 10938176.
- Chinoy ED, Cuellar JA, Huwa KE, et al. Performance of seven consumer sleep-tracking devices compared with polysomnography. Sleep. 2021. PMID 33378539.
- Hersenstichting. The sleep cycle. Accessed 2026.
Disclaimer
Every blood test result includes a professional assessment by a BIG-registered doctor. This article gives general information and is not a substitute for medical advice. For treatment decisions, discuss your results with your GP.
الكاتب