Sleep Stages Explained: Why Losing Two Hours Steals REM for Athletes

Sleep cycles through four distinct stages roughly every 90 to 110 minutes: light NREM sleep (N1 and N2), deep NREM sleep (N3), and REM sleep. Each stage carries its own brain-wave signature and does different work, from physical repair to memory consolidation and emotional processing, which is why cutting a cycle short costs you more than just time.

TL;DR:

• Waking up during deep N3 sleep causes significant grogginess and cognitive impairment lasting up to an hour, making early awakenings particularly disruptive.
• Most deep sleep occurs in the first two cycles of the night, each lasting around 20 to 40 minutes, with REM periods stretching longer toward morning.
• REM sleep, which is concentrated in the second half of the night, is vital for emotional regulation and memory consolidation, with prolonged REM periods becoming more common in the final cycles.
• Shortening sleep or waking early disproportionately reduces REM sleep, which is more sensitive to interruptions than deep N3 sleep.
• Wearable sleep trackers are better at estimating overall sleep patterns and trends over time than providing precise stage details for individual nights.

Table of Contents

• Sleep stages explained: the cycle overview
• NREM stages explained: N1, N2 and N3 in detail
• REM sleep explained: the stage that looks like waking
• Why the stages matter for health, recovery and performance
• What changes your sleep-stage architecture
• How sleep stages are actually measured
• How to protect your deep and REM sleep
• How long each sleep stage typically lasts
• Sleep stages across the lifespan
• Common sleep stage abnormalities worth knowing about
• Using sleep-stage data without overreacting to it
• Track your sleep stages alongside training and recovery
• Sources
• FAQ

Sleep stages explained: the cycle overview

Every night follows a predictable sequence: N1 leads into N2, then N3, back up through N2, and finally into REM before the pattern repeats. This full loop is called a sleep cycle, and it lasts roughly 90 to 110 minutes for most adults, according to NCBI’s StatPearls physiology reference. Across a typical night, you’ll move through four to five of these cycles, though six is not unusual if you’re sleeping a full eight or nine hours.

The stages don’t get equal billing. N2 dominates, accounting for roughly 45% of total sleep time, the biggest single slice of the night. N3 and REM each take up around 20 to 25%, with N1 filling in the small remainder as a brief transitional stage you barely register.

What changes across the night is the ratio between N3 and REM, not the cycle structure itself. Early cycles are heavy with N3, your deepest and most physically restorative sleep. As the night progresses, N3 shrinks and REM periods stretch out, with the longest REM stretches arriving in the final hours before you wake. This is why:

• The first half of the night is when your body does most of its physical repair work
• The second half is when REM dominates, handling more of the cognitive and emotional load
• Waking up early cuts disproportionately into REM, not N3
• Sleeping in past your usual cycle count adds mostly REM-rich time, not deep sleep

Understanding this uneven split matters more than most people realise. It’s the reason a six-hour night doesn’t simply give you 75% of an eight-hour night’s benefits; it can strip out almost all of your REM sleep while leaving your early N3 stages largely intact.

NREM stages explained: N1, N2 and N3 in detail

Non-REM sleep isn’t one uniform state. It’s three progressively deeper stages, each with its own brainwave pattern and physiological job.

N1: the doorway stage. This is the lightest sleep, lasting only about one to seven minutes per cycle. Brain activity slows from the alert wakeful rhythm into slower theta waves, muscle tone loosens, and you can be woken by the smallest noise or a twitch of your own leg. Most people who claim they “never really sleep” are underestimating how much time they spend cycling briefly through N1 without noticing.

N2: the workhorse stage. This is where sleep spindles and K-complexes appear on an EEG trace, brief bursts of fast brain-wave activity that show up nowhere else in the sleep cycle. Spindle density has been linked to memory consolidation, particularly for procedural skills and factual recall, according to StatPearls. Heart rate and body temperature both drop slightly here. N2 takes up the largest share of any single stage across the night.

N3: slow-wave sleep. Also called deep sleep, N3 is dominated by slow delta waves and is the stage most tied to physical restoration. Growth hormone release peaks here, tissue repair accelerates, and immune function gets a measurable boost. It’s also the stage where parasomnias like sleepwalking and night terrors tend to occur, because the brain is deeply asleep while parts of the motor system remain oddly active.

N3 makes up roughly 20 to 25% of total sleep in a healthy adult, concentrated heavily in the first two cycles of the night, according to NCBI’s sleep physiology overview.

Waking directly out of N3 produces a distinctive form of grogginess called sleep inertia, which can blunt reaction time and decision making for 30 to 60 minutes afterwards. This is worth knowing if you’ve ever been jolted awake by an alarm mid deep-sleep and felt genuinely impaired for the first part of your morning, that’s not laziness, it’s a measurable physiological lag.

A few features distinguish the three NREM stages at a glance:

• N1: theta waves, easy arousal, brief duration, functions mainly as a transition
• N2: sleep spindles and K-complexes, moderate arousal threshold, largest share of total sleep
• N3: delta waves, hardest to wake from, tied to physical repair and immune health

REM sleep explained: the stage that looks like waking

REM sleep is the strangest stage on an EEG trace, brain activity resembles wakefulness far more than it resembles deep sleep, while the body is almost completely paralysed. This mismatch, called REM atonia, prevents you from acting out dreams; the eyes dart rapidly beneath closed lids, giving the stage its name.

The first REM period of the night arrives around 90 minutes after you fall asleep and tends to be short, often just a few minutes. Each subsequent REM period lengthens, with the final one before waking sometimes stretching past 20 or 30 minutes. This concentration toward the morning hours explains why the dreams you remember most vividly are usually the ones just before your alarm goes off.

REM does distinct cognitive work that NREM sleep doesn’t replicate:

• Emotional processing: REM appears to help regulate and defuse the intensity of emotionally charged memories overnight
• Certain types of memory consolidation, particularly those involving complex or emotionally salient material
• Creative problem solving, with some evidence linking REM to improved insight and pattern recognition the next day

Because REM is backloaded into the second half of the night, losing sleep at the tail end, hitting snooze repeatedly or waking two hours early, costs you disproportionately more REM than deep sleep. Given REM’s role in mood regulation, this is a plausible mechanism behind why chronic early waking correlates so strongly with irritability and low mood, a pattern the National Institute of Neurological Disorders and Stroke points to when describing how sleep architecture shifts across the night.

Why the stages matter for health, recovery and performance

Each stage maps onto a different outcome you actually care about. N3 drives physical repair and immune function, which is why athletes and anyone recovering from illness lean so heavily on getting enough deep sleep. REM and N2, by contrast, are more closely tied to memory consolidation and emotional regulation, the stuff that determines how sharp and steady you feel the next day rather than how repaired your muscles are.

Shortfalls in either direction produce recognisable symptoms. Chronic N3 loss tends to show up as poor physical recovery, slower adaptation to training, and a weaker immune response. Chronic REM loss shows up more as cognitive fog, mood instability, and difficulty processing emotionally taxing events. If you’ve ever noticed that a short but intense sleep debt leaves you clumsy and slow rather than upset, and a run of early wake-ups leaves you snappish rather than physically wrecked, that split isn’t a coincidence.

Age changes this picture considerably. Children spend a notably higher proportion of the night in N3, which tracks with the sheer volume of physical growth and development happening during childhood. Older adults typically see a decline in slow-wave sleep and more fragmented N2, changes that are largely a normal part of ageing rather than a red flag on their own. That said, if fragmentation becomes severe, with frequent full awakenings or a near-total absence of deep sleep, it’s worth raising with a clinician rather than assuming it’s simply age catching up.

What changes your sleep-stage architecture

Several everyday factors reshape how much time you spend in each stage, often without you noticing until the pattern repeats for weeks.

• Sleep restriction: cutting a night short hits REM hardest, since REM is concentrated toward the morning hours you’re skipping.
• Alcohol: it can help you fall asleep faster but suppresses REM in the first half of the night and fragments sleep later on, once the sedative effect wears off, a pattern covered in more depth in Heala’s breakdown of alcohol’s fingerprint on your sleep graph.
• Room temperature: REM blunts your body’s normal thermoregulation, so a bedroom that’s too warm disrupts REM continuity more than it disrupts N3, according to NINDS.
• Medications and stimulants: many antidepressants suppress REM, while stimulants taken too late in the day delay sleep onset and shrink the number of full cycles you complete.
• Sleep disorders and shift work: obstructive sleep apnoea fragments N3 with repeated micro-arousals, and irregular shift schedules disrupt the circadian signal that normally times REM toward the morning.

How sleep stages are actually measured

Polysomnography is the clinical gold standard, and it earns that status by measuring more than one signal at once. A full sleep study records brain activity via EEG, eye movement via EOG, and muscle activity via EMG, plus respiratory and cardiac data, all scored together to classify each 30-second window into a specific stage, per NHLBI’s overview of sleep phases.

Consumer wearables can’t do this. Most infer sleep stages from movement (actigraphy) and heart-rate variability, proxies that correlate with stage transitions but don’t measure brain activity directly. This has a real practical consequence: wearables commonly misclassify quiet wakefulness as light sleep, since lying still with your eyes closed looks similar to a device tracking motion and heart rate alone, a limitation NHLBI notes explicitly.

None of that makes wearable data useless. It just changes what you should trust it for.

• Total sleep duration and bedtime consistency: generally reliable
• Broad light-sleep-versus-deep-sleep trends over weeks: reasonably useful
• Exact minute counts for a single night’s N2 or REM: treat with scepticism
• Night-to-night comparisons on the same device: more meaningful than comparing across different brands, since algorithms vary between manufacturers

Pro Tip: Stick to one device for your own tracking rather than switching between brands. Because staging algorithms differ, jumping between a fitness watch and a ring tracker can make your sleep look wildly inconsistent when nothing about your actual sleep has changed.

How to protect your deep and REM sleep

Improving your sleep-stage balance isn’t about hacking one stage in isolation, it’s mostly about giving your body the conditions and time it needs to run its natural sequence uninterrupted.

1. Keep a consistent sleep and wake time, including weekends. Irregular timing disrupts the circadian signal that governs when REM-heavy cycles occur, pushing them later or shortening them.
2. Protect your full sleep opportunity. Since REM concentrates in the final cycles, waking two hours early doesn’t just remove two hours evenly, it disproportionately strips out REM.
3. Cut off alcohol several hours before bed. REM suppression in the early night and fragmentation later are both dose-dependent, so timing your last drink earlier reduces the disruption.
4. Keep the bedroom cool. Because REM blunts your normal temperature regulation, a room in the high teens Celsius protects REM continuity better than a warm one.
5. Watch late caffeine and stimulant use. Even six hours before bed, stimulants can delay sleep onset enough to shave a full cycle off the night.
6. See a clinician if fragmentation looks severe or you suspect a disorder like sleep apnoea or REM behaviour disorder, particularly if a partner reports repeated pauses in breathing, loud snoring, or you acting out dreams physically.

Pro Tip: If you only have time to fix one habit, fix your wake time first. A stable wake time anchors the whole cycle sequence, which does more for REM protection over a month than any single bedtime ritual.

How long each sleep stage typically lasts

Within a single 90 to 110 minute cycle, the four stages don’t take equal shares of time, and that share shifts depending on where you are in the night.

N1 is brief by design, usually just one to seven minutes, acting purely as the transition out of wakefulness. N2 tends to run the longest within any given cycle, often 20 minutes or more once you’re past the first cycle of the night. N3 varies enormously by cycle position: in the first cycle it might last 20 to 40 minutes, but by the third or fourth cycle it can shrink to almost nothing as REM takes over the back half of the night. REM itself starts short, often under 10 minutes in the first cycle, and lengthens with each subsequent cycle, sometimes exceeding 30 minutes by the final cycle before waking.

This is why looking at a single night’s stage breakdown in isolation can be misleading. A night with a lot of early waking will show a chunk of solid N3 but a REM percentage far below your usual average, not because anything went wrong with your deep sleep, but because you simply didn’t complete enough cycles to reach the REM-heavy stretch. Consistency across nights, not any single figure, is what tells you whether your architecture is actually healthy.

Sleep stages across the lifespan

Sleep architecture isn’t fixed at any age; it shifts predictably as you move from infancy through to later life, and each phase has its own version of “normal.”

Infants spend a strikingly high proportion of sleep in active sleep, the developmental precursor to REM, which is thought to support the rapid brain growth happening in the first year. Their cycles are also much shorter, often 50 to 60 minutes rather than the 90 to 110 minute adult cycle, meaning infants pass through far more transitions per night. Children and adolescents carry a much higher share of N3 than adults, which tracks with the physical growth and tissue repair demands of that stage of life.

By adulthood, the proportions settle into the familiar pattern of roughly 45% N2, 20 to 25% N3, and 20 to 25% REM. Older adults typically see a further shift: slow-wave sleep declines, awakenings become more frequent, and N2 takes up a larger relative share as N3 shrinks. This is largely a normal physiological change rather than a disorder on its own, though it does help explain why older adults often report lighter, more easily disrupted sleep even without an underlying condition. The practical takeaway is that comparing your sleep stages against a fixed “ideal” percentage ignores how much that ideal moves with age.

Common sleep stage abnormalities worth knowing about

Two patterns come up often enough in sleep medicine that they’re worth understanding even outside a clinical context.

REM behaviour disorder involves a failure of the normal muscle paralysis that should accompany REM sleep. Instead of lying still while dreaming, people with this disorder can physically act out dream content, sometimes violently enough to injure themselves or a bed partner. It’s a distinct condition from ordinary sleep talking or occasional twitching, and it’s associated with neurological changes that in some cases precede other conditions by years, which is why persistent, dream-enacting movement during sleep warrants a clinical conversation rather than a home fix.

Fragmented N3 sleep is more common and usually less alarming in isolation, but it still carries real consequences. Repeated brief arousals during deep sleep, often driven by obstructive sleep apnoea, noise, or pain, prevent the brain from completing full slow-wave cycles even when total sleep time looks adequate on paper. The result is someone who technically slept eight hours but wakes feeling as though they barely slept, because the restorative benefit of N3 depends on sustained, uninterrupted stretches rather than total minutes accumulated in fragments.

Both patterns illustrate the same principle: how sleep is structured matters as much as how long it lasts.

Using sleep-stage data without overreacting to it

Sleep-stage estimates are trend tools, not verdicts on a single bad night. The most useful way to read them is alongside how you actually feel, your subjective energy, soreness, and mental clarity, rather than treating a wearable’s REM percentage as gospel.

When a tracker shows reduced N3 or heavily fragmented REM after a rough night, that’s a reasonable signal to ease off training intensity or push a hard session back a day, not a reason to panic. What matters far more is the pattern across two or three weeks. A single night of poor deep sleep after a late flight or a stressful evening tells you almost nothing on its own; a consistent multi-week dip alongside rising resting heart rate or falling heart rate variability tells you something worth acting on.

The trap most people fall into is reacting to daily fluctuations as if they were diagnoses. Sleep-stage percentages vary meaningfully between individuals and between consecutive nights for the same person, even when nothing is actually wrong, a point StatPearls makes clear when describing normal variation. Treat the numbers as one input among several, not the whole picture.

— Kerem

Track your sleep stages alongside training and recovery

Heala pulls sleep data straight from the wearable you already use, Fitbit, WHOOP, Oura, Apple Health, or Health Connect, and lines it up against your training load, heart rate variability, and daily readiness in one place, rather than leaving you to cross-reference three separate apps.

The platform can flag when a run of reduced deep sleep or fragmented REM coincides with flat recovery scores, giving you a clearer trend than any single night’s stage breakdown could. That context matters more than a raw percentage on its own, and it’s built to work whether or not you’re wearing a strap every night. If you want to see how your own sleep-stage trends line up with your training and recovery, the sleep tracking feature is a good place to start, and you can explore the full app to see how sleep, nutrition, and workouts connect in your daily plan.

This article is general information, not a substitute for advice from a qualified doctor. Consult a qualified healthcare professional about your own circumstances before acting on anything here.

Sources

• Physiology, Sleep Stages - StatPearls - NCBI Bookshelf
• Understanding Sleep (Brain Basics) — NINDS / NIH
• How sleep works — Sleep phases and stages | NHLBI, NIH

FAQ

How much of each sleep stage should you get?

In a typical adult night, roughly 45% is spent in N2, 20 to 25% in N3, and 20 to 25% in REM, with N1 making up a small remainder, according to StatPearls.

Is it better to get more REM or more deep sleep?

Neither outranks the other; N3 drives physical repair and immune function, while REM supports emotional processing and memory consolidation, so a healthy night needs both rather than a surplus of one.

Is two hours of REM sleep good or bad?

Two hours of REM is within a normal range for a full night’s sleep, since REM typically makes up 20 to 25% of total sleep time, but context matters more than the raw figure, check it against your usual pattern rather than a fixed target.

Which sleep stage is worst to wake up from?

Waking directly out of N3, deep slow-wave sleep, tends to produce the most sleep inertia, often leaving you groggy and slower to think clearly for 30 to 60 minutes afterwards.

Can a wearable accurately tell me my sleep stages?

Wearables estimate stages using movement and heart-rate proxies rather than brain activity, so they’re useful for spotting trends over weeks but shouldn’t be treated as a clinical-grade breakdown of a single night, a limitation NHLBI notes directly.

Recommended

• Alcohol’s exact fingerprint on your sleep graph
• Overtraining is rare. Under-recovering is everywhere
• Naps: what 20 minutes buys and 90 minutes costs