The Science of Sleep: What Happens in Your Brain Each Night
Sleep isn’t a shutdown—it’s a structured sequence of cycles, stages, and shifting brain states. Here’s what actually runs overnight, and why it matters.
By TheMurrow Editorial
January 19, 2026
Key Points
- 1Know the architecture: adult sleep cycles through NREM and REM about 4–6 times per night, reshaping what “eight hours” really means.
- 2Track what matters: prioritize consistency and continuity, because fragmentation boosts lighter stages like N1 and can leave you unrefreshed.
- 3Interpret wearables cautiously: without EEG, stage labels are estimates—use trends as clues and pursue PSG if symptoms persist.
Most nights, you don’t simply “go to sleep.” You run a program.
It starts quietly: the mind loosens its grip on the day, muscles slacken, thoughts blur at the edges. Then the brain shifts into a repeating pattern—an internal schedule that looks less like a single plunge into darkness and more like a series of timed acts.
The surprise is how structured the night really is. Adult sleep moves through two major states—non-REM (NREM) and REM sleep—cycling repeatedly. According to the U.S. National Heart, Lung, and Blood Institute (NHLBI), a typical night includes about 4–6 cycles, each lasting roughly 80–100 minutes, though the timing varies by person and changes across the night.
4–6
Typical adult nights include about 4–6 sleep cycles, repeating NREM and REM in a structured overnight program (NHLBI).
80–100 min
Each sleep cycle often lasts roughly 80–100 minutes, though timing varies by person and shifts across the night (NHLBI).
If you’ve ever woken near morning and felt as if you’d been dreaming more, you weren’t imagining it. The Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) notes a key pattern: deep NREM sleep dominates earlier, while REM becomes more frequent later, with the final hours skewing toward lighter sleep and REM. The arc of the night is one reason why “eight hours” is not a monolith. When you sleep matters almost as much as how long.
Sleep isn’t a shutdown. It’s a sequence.
— — TheMurrow Editorial
The night’s blueprint: cycles, stages, and a shifting balance
A full night of adult sleep alternates between NREM (a set of stages) and REM (the dream-rich state most people recognize). The overall structure is remarkably consistent across healthy adults, even if the details vary.
The cycle you repeat 4–6 times
NHLBI’s public guidance describes sleep as cycling through NREM and REM repeatedly, with a typical night containing ~4–6 cycles of ~80–100 minutes each. Early in the night, those cycles are weighted toward deeper NREM sleep. Later, REM expands, often arriving in longer stretches.
That timing has everyday consequences. If someone sleeps from 10 p.m. to 6 a.m., they’re not getting the same “mix” of sleep as someone who sleeps from 2 a.m. to 10 a.m.—even if the total hours match—because the brain’s architecture changes across the night.
That timing has everyday consequences. If someone sleeps from 10 p.m. to 6 a.m., they’re not getting the same “mix” of sleep as someone who sleeps from 2 a.m. to 10 a.m.—even if the total hours match—because the brain’s architecture changes across the night.
“Normal” isn’t one number
Sleep labs often talk about stage percentages, but the research record is careful about what “normal” means. Different studies use different cohorts, age ranges, and scoring practices; clinical populations can look different from healthy volunteers.
Still, approximate adult proportions provide a useful map:
- N1: ~5% of the night (a light transitional stage)
- N2: ~45–60% (often the largest share)
- N3: ~10–20% (slow-wave or “deep” sleep)
- REM: ~20–25%
These figures appear across commonly cited summaries in clinical and research contexts, including reviews hosted on PubMed Central and clinical references such as Medscape. Research also stresses that REM can be lower in some datasets, especially in clinical contexts, reinforcing the point: stage percentages are guides, not verdicts.
Still, approximate adult proportions provide a useful map:
- N1: ~5% of the night (a light transitional stage)
- N2: ~45–60% (often the largest share)
- N3: ~10–20% (slow-wave or “deep” sleep)
- REM: ~20–25%
These figures appear across commonly cited summaries in clinical and research contexts, including reviews hosted on PubMed Central and clinical references such as Medscape. Research also stresses that REM can be lower in some datasets, especially in clinical contexts, reinforcing the point: stage percentages are guides, not verdicts.
45–60%
N2 commonly makes up roughly 45–60% of adult sleep—often the largest share—despite being mislabeled as “just light sleep.”
Eight hours isn’t one thing—it’s a changing composition across the night.
— — TheMurrow Editorial
How scientists know what stage you’re in (and why your wearable struggles)
Sleep staging is not guesswork. In the lab, it’s a measurement problem—solved with a lot of sensors and strict rules.
Polysomnography: the gold standard
The clinical and research gold standard is polysomnography (PSG). NHLBI describes how sleep stages are classified using signals that track:
- EEG (brain electrical activity)
- EOG (eye movements)
- EMG (muscle tone)
- plus cardiorespiratory monitoring (breathing, oxygen levels, heart rate)
Those channels matter because sleep is not only behavior (stillness, closed eyes). Sleep is also brain-state. Two people can lie equally still while their brains show very different signatures.
- EEG (brain electrical activity)
- EOG (eye movements)
- EMG (muscle tone)
- plus cardiorespiratory monitoring (breathing, oxygen levels, heart rate)
Those channels matter because sleep is not only behavior (stillness, closed eyes). Sleep is also brain-state. Two people can lie equally still while their brains show very different signatures.
The 30-second reality of sleep scoring
In clinical practice, sleep is typically scored in discrete windows called epochs, commonly 30 seconds per epoch, using standardized criteria. The American Academy of Sleep Medicine (AASM) provides the core technical rulebook: the AASM Scoring Manual, updated as recently as Version 3 (released February 2023), with implementation required by the end of 2023.
That detail matters because it reveals the granularity of the craft. Sleep isn’t assessed as one continuous stream. It’s parsed, labeled, and audited in half-minute slices.
That detail matters because it reveals the granularity of the craft. Sleep isn’t assessed as one continuous stream. It’s parsed, labeled, and audited in half-minute slices.
30 seconds
In clinical scoring, sleep is typically labeled in 30-second epochs, reflecting how granular—and rule-bound—professional staging really is.
Why consumer “sleep stages” are partial truths
Many wearables estimate sleep stages primarily from movement and heart-related signals rather than EEG. Mainstream clinical commentary often frames these tools as helpful for tracking routine and consistency, but limited for precise stage identification compared with PSG.
A practical way to read wearable data: treat the trend as information and the exact labels as approximations. If your device reports unusually low “deep sleep” one night, the useful question may be whether you slept poorly overall, drank alcohol late, or were stressed—not whether the number itself is an exact measurement of N3.
A practical way to read wearable data: treat the trend as information and the exact labels as approximations. If your device reports unusually low “deep sleep” one night, the useful question may be whether you slept poorly overall, drank alcohol late, or were stressed—not whether the number itself is an exact measurement of N3.
A smartwatch can notice your stillness. A sleep lab measures your brain.
— — TheMurrow Editorial
N1: the borderland where “I wasn’t asleep” often is
N1 is the first stage of NREM sleep, often described as a transitional state between wakefulness and sleep. NHLBI characterizes it as the lightest stage, where people can be easily awakened.
Why N1 feels like limbo
Many people have had the experience: you’re drifting, you feel awake, and then a sudden jolt—your body twitches, a thought snaps you back, or you realize minutes passed. N1 is that boundary zone. It’s also where subjective experience can mislead. Someone can insist they were awake while PSG scoring shows they entered N1.
That mismatch is not a moral failure or “overthinking.” It’s a reminder that consciousness is not a perfect self-monitoring tool. Sleep onset can be subtle.
That mismatch is not a moral failure or “overthinking.” It’s a reminder that consciousness is not a perfect self-monitoring tool. Sleep onset can be subtle.
When N1 becomes a clue
Sleep labs also treat elevated N1 as a potential marker of fragmentation. Frequent arousals—whether from stress, discomfort, breathing issues, or other disturbances—can increase time spent in lighter stages. A night with more N1 can feel less restorative, even if total sleep time looks adequate on paper.
For readers, the implication is practical: if you routinely feel unrefreshed, the problem may not be that you “didn’t get enough hours.” The issue may be how continuous your sleep was and how often you were nudged back toward the surface.
For readers, the implication is practical: if you routinely feel unrefreshed, the problem may not be that you “didn’t get enough hours.” The issue may be how continuous your sleep was and how often you were nudged back toward the surface.
N2: the workhorse stage where the brain gates the world
If sleep were a budget, N2 would be the largest line item. Clinical summaries commonly estimate N2 at roughly 45–60% of adult sleep, making it the stage many people undervalue simply because it lacks the mystique of “deep sleep” or the drama of dreaming.
Spindles and K-complexes: the signatures of N2
N2 is defined by distinctive EEG features, including sleep spindles (brief bursts of rhythmic activity) and K-complexes. Research literature links these features to how the brain processes information during sleep—often discussed in terms of learning, sensory gating, and memory-related functions.
A 2023 review indexed on PubMed (PubMed ID: 37597610) highlights ongoing scientific interest in sleep spindles and their relationship to cognitive processes. Researchers are still debating precise roles and mechanisms, but the broad takeaway is sturdy: N2 is active, not empty.
A 2023 review indexed on PubMed (PubMed ID: 37597610) highlights ongoing scientific interest in sleep spindles and their relationship to cognitive processes. Researchers are still debating precise roles and mechanisms, but the broad takeaway is sturdy: N2 is active, not empty.
Real-world example: the “light sleep” misunderstanding
Consider a common scenario. A professional checks a wearable report and sees “mostly light sleep.” Anxiety follows: “No wonder I’m tired.” The problem is semantic. Many devices lump N1 and N2 into “light,” but N2 is supposed to be a large share of the night in healthy adults.
That doesn’t mean every “light sleep” report is fine. It means interpretation matters. A sleep lab distinguishes between N1 (a fragile edge) and N2 (a stable stage with specific brain features). A consumer app often does not.
That doesn’t mean every “light sleep” report is fine. It means interpretation matters. A sleep lab distinguishes between N1 (a fragile edge) and N2 (a stable stage with specific brain features). A consumer app often does not.
N3: deep sleep and the night’s heavy engineering
N3—often called slow-wave sleep or “deep sleep”—gets the reputation as the night’s most restorative block. NHLBI notes that deeper NREM sleep tends to occur more in the first part of the night, aligning with the NICHD description of a night that becomes lighter toward morning.
Harder to wake, quieter responsiveness
The National Institute of Neurological Disorders and Stroke (NINDS) describes deeper NREM sleep as a state where arousal responsiveness drops. Anyone who has tried to rouse a child—or an adult—during deep sleep recognizes the difference. The brain is less receptive to external prompts.
That reduced responsiveness is not merely inconvenience. It signals that the brain is running a different mode, prioritizing internal processes over environmental monitoring.
That reduced responsiveness is not merely inconvenience. It signals that the brain is running a different mode, prioritizing internal processes over environmental monitoring.
A major idea: synaptic homeostasis
One influential framework in sleep science is the synaptic homeostasis hypothesis: the notion that waking life broadly strengthens synapses, and sleep—particularly slow-wave activity—helps downscale and renormalize them. The theory is not a slogan; it’s a lens used to interpret why slow-wave activity might be so central to brain function.
Scientists continue to test and refine this idea, and competing theories exist. Still, the core intuition resonates with lived experience: after long periods of learning, stimulation, and stress, the brain may need a systematic reset.
Scientists continue to test and refine this idea, and competing theories exist. Still, the core intuition resonates with lived experience: after long periods of learning, stimulation, and stress, the brain may need a systematic reset.
REM: the later-night state that changes the story you tell about sleep
REM sleep tends to rise as the night progresses. NICHD notes that sleep cycles shift toward more REM later, with lighter stages and REM dominating closer to morning. That pattern changes what you remember.
Why morning awakenings feel dreamier
People often recall dreams when awakened during REM, and the final hours of sleep contain more REM. That’s one reason early alarms can feel psychologically jarring: you’re more likely to interrupt a REM-rich portion of the night.
REM’s prominence later also explains why “sleeping in” can sometimes feel disproportionately helpful. Even an extra hour may capture a REM-heavy window that you’d otherwise cut off.
REM’s prominence later also explains why “sleeping in” can sometimes feel disproportionately helpful. Even an extra hour may capture a REM-heavy window that you’d otherwise cut off.
A practical implication for schedules
A consistent schedule doesn’t just stabilize total sleep time. It stabilizes where your awakenings land in the architecture of the night. Someone with variable bedtimes can end up truncating early-night N3 on one day and late-night REM on another, producing different kinds of tiredness.
For shift workers or new parents, this is less a lifestyle choice than a reality. The point isn’t moralizing; it’s understanding why fragmented nights can feel uniquely disorienting. You’re not only losing minutes—you’re losing specific phases.
For shift workers or new parents, this is less a lifestyle choice than a reality. The point isn’t moralizing; it’s understanding why fragmented nights can feel uniquely disorienting. You’re not only losing minutes—you’re losing specific phases.
The politics of “deep sleep” online: what your data can and can’t tell you
Sleep has become quantified, optimized, and discussed in the language of scores. That shift has benefits: people pay attention. The risk is that attention hardens into mistaken certainty.
What stage percentages can obscure
Stage percentages are averages. They vary by age, health, and measurement method. The research record is explicit that “normal” is not one fixed template; different samples yield different distributions, and some clinical datasets report lower REM.
A wearable that presents stage percentages as precise can invite overinterpretation. Without EEG, classification becomes inference. Even with EEG, sleep is scored in 30-second epochs, meaning a “stage” is already a label placed on a short slice of complex biology.
A wearable that presents stage percentages as precise can invite overinterpretation. Without EEG, classification becomes inference. Even with EEG, sleep is scored in 30-second epochs, meaning a “stage” is already a label placed on a short slice of complex biology.
Multiple perspectives: skeptics and pragmatists
One camp argues consumer stage tracking is too noisy to be meaningful. Another argues that imperfection doesn’t equal uselessness: trend lines can still reveal patterns—late caffeine, alcohol, inconsistent bedtimes, stress—especially when paired with daytime symptoms.
Both views can be true. The sensible middle ground looks like this:
- Use wearables to track regularity, duration, and broad sleep disruption
- Treat specific stage readouts as estimates, not diagnoses
- If persistent fatigue, snoring, or frequent awakenings appear, consider clinical evaluation with PSG
Both views can be true. The sensible middle ground looks like this:
- Use wearables to track regularity, duration, and broad sleep disruption
- Treat specific stage readouts as estimates, not diagnoses
- If persistent fatigue, snoring, or frequent awakenings appear, consider clinical evaluation with PSG
Middle-ground rule for sleep trackers
Use wearables to track regularity and disruption.
Treat stage labels as estimates.
Escalate persistent symptoms (fatigue, snoring, awakenings) to PSG-based clinical evaluation.
Treat stage labels as estimates.
Escalate persistent symptoms (fatigue, snoring, awakenings) to PSG-based clinical evaluation.
Practical takeaways: how to think about your night like a sleep clinician
Sleep science can feel abstract until you apply it to your own routine. The goal isn’t to micromanage the night. The goal is to stop asking the wrong questions.
What to watch for (and what to stop obsessing over)
Focus on:
- Consistency: regular bed and wake times stabilize the night’s shifting balance of NREM and REM
- Continuity: fewer awakenings often means less time stuck in N1 and more time progressing through cycles
- Context: stage percentages vary; compare yourself to yourself over time, not to a single “ideal” pie chart
Obsessing over a single night’s “deep sleep minutes” misses the bigger structure. A typical night contains multiple cycles; one disrupted cycle doesn’t define you.
- Consistency: regular bed and wake times stabilize the night’s shifting balance of NREM and REM
- Continuity: fewer awakenings often means less time stuck in N1 and more time progressing through cycles
- Context: stage percentages vary; compare yourself to yourself over time, not to a single “ideal” pie chart
Obsessing over a single night’s “deep sleep minutes” misses the bigger structure. A typical night contains multiple cycles; one disrupted cycle doesn’t define you.
Clinician-style focus areas
- ✓Prioritize consistency in bed and wake times
- ✓Protect continuity by reducing awakenings where possible
- ✓Use context: compare trends over time, not a single-night stage chart
Case study: the early alarm problem
Imagine a reader who sleeps 11:30 p.m. to 6:00 a.m. on weekdays and 2:00 a.m. to 10:00 a.m. on weekends. Total hours may average out, but the brain’s architecture won’t. Weekday sleep truncates late-night REM; weekend sleep shifts the entire program later.
The result can feel like persistent jet lag. NICHD’s description of increasing REM later in the night makes this pattern legible: you’re repeatedly cutting off a REM-rich portion, then trying to pay it back in bulk.
The result can feel like persistent jet lag. NICHD’s description of increasing REM later in the night makes this pattern legible: you’re repeatedly cutting off a REM-rich portion, then trying to pay it back in bulk.
When to seek more than self-tracking
If sleep feels persistently unrefreshing despite adequate time in bed, or if a partner reports significant snoring or pauses in breathing, a clinical workup matters. PSG is designed for precisely that: measuring brain state and physiology together rather than inferring sleep from motion.
A wearable can suggest a pattern. A sleep lab can test a hypothesis.
A wearable can suggest a pattern. A sleep lab can test a hypothesis.
Key Insight
If you’re consistently unrefreshed—or there’s loud snoring or breathing pauses—treat tracker data as a clue and consider PSG to test what’s actually happening.
A night well-lived is a night well-structured
The most bracing lesson in modern sleep science is how little “sleep” resembles a single thing. Your brain runs a schedule: N1’s threshold, N2’s spindled stability, N3’s slow-wave depth early in the night, and REM’s expanding presence toward morning. The program repeats 4–6 times, in cycles of roughly 80–100 minutes, while the balance shifts hour by hour.
Understanding that architecture changes the questions worth asking. Instead of chasing a perfect stage chart, aim for conditions that allow cycles to unfold with minimal interruption. Instead of treating morning grogginess as a personal failure, recognize it may be the predictable result of waking from a REM-heavy stretch. And instead of letting consumer metrics dictate your mood, use them the way clinicians use measurements: as clues, not commandments.
Sleep is not a blank space between days. It’s a nightly sequence your brain depends on—whether you remember it or not.
Understanding that architecture changes the questions worth asking. Instead of chasing a perfect stage chart, aim for conditions that allow cycles to unfold with minimal interruption. Instead of treating morning grogginess as a personal failure, recognize it may be the predictable result of waking from a REM-heavy stretch. And instead of letting consumer metrics dictate your mood, use them the way clinicians use measurements: as clues, not commandments.
Sleep is not a blank space between days. It’s a nightly sequence your brain depends on—whether you remember it or not.
Sleep is not a blank space between days. It’s a nightly sequence your brain depends on—whether you remember it or not.
— — TheMurrow Editorial
T
About the Author
TheMurrow Editorial is a writer for TheMurrow covering science.
Frequently Asked Questions
How many sleep cycles should I get per night?
Public guidance from NHLBI describes a typical adult night as containing about 4–6 cycles, each roughly 80–100 minutes. The exact number varies by person and by how long you sleep. Shorter nights reduce the number of full cycles you complete, which can change how much late-night REM you get.
Why do I dream more in the morning?
NICHD notes that REM sleep becomes more frequent later in the night, with sleep skewing toward lighter stages and REM near morning. Because dreams are often recalled when waking from REM, awakenings closer to your alarm tend to feel more dream-heavy than awakenings earlier in the night.
What are “normal” percentages for deep sleep and REM?
Approximate adult figures often cited include N3 (deep sleep) around 10–20% and REM around 20–25%, with N2 around 45–60% and N1 around 5%. Research cautions that “normal” is not one universal standard; results vary by age, health, and how sleep is measured and scored.
Why does my wearable say I get almost no deep sleep?
Many wearables infer stages from movement and heart-related signals rather than EEG, so stage precision is limited compared with polysomnography. Treat stage numbers as estimates and look for trends over time. If you feel consistently unrefreshed, consider discussing symptoms with a clinician rather than relying on a device label.
What’s the difference between N1 and N2 “light sleep”?
N1 is a transition from wake to sleep and is easy to disrupt; NHLBI describes it as the lightest stage. N2 is typically the largest share of adult sleep (often 45–60%) and has distinct brain signatures such as sleep spindles and K-complexes. Many consumer apps bundle both into “light,” which can confuse interpretation.
How do sleep labs actually determine my sleep stage?
Sleep labs use polysomnography (PSG), which includes EEG (brain activity), EOG (eye movements), and EMG (muscle tone), along with cardiorespiratory measures. Sleep is scored in epochs, commonly 30 seconds, using standards from the AASM Scoring Manual (updated with Version 3 released in February 2023).
