If you own an Apple Watch you have seen "Core" listed as a sleep stage in the Health app, sitting alongside Deep and REM with no explanation of what it is. Core sleep is Apple's name for light sleep — the combined N1 and N2 stages that every other tracker also measures. The label is Apple-specific. The physiology is not.
The Verdict
What core sleep is physiologically
N1 is the transition into sleep. It lasts a few minutes, muscle tone drops, and you can be woken from it without ever believing you were asleep. N2 is the substantial one. It occupies roughly half of a normal adult night and carries two distinctive brain signatures: sleep spindles and K-complexes.
Sleep spindles are short bursts of 11–16 Hz activity generated by the thalamus. They are not filler. Spindle density tracks with overnight memory consolidation, and people who produce more spindles retain more of what they learned the day before. K-complexes appear to suppress arousal from external noise, which is part of why N2 keeps you asleep through a passing siren. During N2 core body temperature falls, heart rate slows, and breathing regularises.
Calling this "light" sleep is a description of arousal threshold, not of value. You are easier to wake from N2 than from N3. That is all the word means.
How much core sleep is normal
| Stage | What it is | Share of a normal adult night | Notes |
|---|---|---|---|
| Awake | Brief arousals between cycles | 0–30 min after sleep onset | Some awake time is normal. Over 30 minutes points to fragmentation, not a tracking error. |
| Core (N1 + N2) | Light-to-moderate stage sleep — the bulk of the night | 50–60% of total sleep | N2 sleep spindles do real memory work. Core is not idle time between the "good" stages. |
| Deep (N3) | Slow-wave sleep — growth hormone, glymphatic clearance | 13–23% of total sleep | Front-loaded. Most of it lands in the first two cycles of the night. |
| REM | Dream sleep — emotional and procedural memory | 20–25% of total sleep | Back-loaded. The last two cycles before waking carry most of it. |
On a 7.5-hour night, 50–60% core works out to roughly 3 hours 45 minutes to 4 hours 30 minutes. Age shifts the split. Deep sleep falls about 2% per decade from the twenties onward, and the minutes it loses mostly reappear as core. A healthy 55-year-old often shows 60–65% core with 10–14% deep, and that is the expected physiology for that age rather than a habit failure. Comparing your split against a 25-year-old's is comparing against a different body.
Sex differences in stage percentages are small in adults and smaller than the night-to-night variation within one person, so there is no separate male and female target worth publishing. Women do report more fragmented sleep in the luteal phase and through perimenopause, which raises awake time and core share without changing what a healthy split should look like.
Why Apple calls it "core" and everyone else calls it "light"
| Brand | Label for N1 + N2 | Stages reported | Notes |
|---|---|---|---|
| Apple Watch | Core | Awake, REM, Core, Deep | The only major brand using the word "Core" for N1 + N2. |
| Fitbit | Light | Awake, REM, Light, Deep | Same four-bucket model, different label. |
| Garmin | Light | Awake, REM, Light, Deep | Adds a separate restlessness measure alongside the stages. |
| Oura | Light | Awake, REM, Light, Deep | Reports latency and efficiency next to the stage split. |
| Whoop | Light | Awake, REM, Light, SWS | Uses "SWS" (slow-wave sleep) rather than "Deep". |
The table above is listed alphabetically by brand. Apple's naming choice is defensible — N1 and N2 genuinely form the core of each cycle — but it creates a translation problem when you read advice written for any other device. Any article telling you to reduce your light sleep is talking about your core sleep number, and the advice is usually wrong anyway, for reasons covered below.
How it is measured, and how wrong it can be
A sleep lab scores stages from polysomnography: EEG electrodes reading cortical activity, plus eye movement and chin muscle tone. Stage boundaries are defined by brainwave patterns in 30-second epochs. Your watch has none of that. It has an optical heart-rate sensor, an accelerometer, and a wrist temperature sensor, and a model that infers stages from those three signals.
That inference is good at some jobs and poor at others. Sleep-versus-wake classification is strong, and total sleep time typically lands within 10–20 minutes of lab scoring. Stage classification is the weak point, generally 60–80% agreement against expert scoring, and core is where most of the error accumulates — it is the default bucket the model falls back to when the signal is ambiguous. Deep sleep misclassified as core, and REM misclassified as core, both inflate the core number without anything changing in your brain.
How to act on the number — and when not to
Do not react to one night. Night-to-night stage variation in the same healthy person routinely swings 20–30 minutes per stage, and wearing two devices simultaneously produces splits that differ by about that much. A single high-core night carries no information.
Read a 7-day and 30-day average instead. If core sits between 50% and 60% and deep and REM are both in range, there is nothing to fix. If core has climbed 10 percentage points versus your own 30-day baseline, work the causes in order: alcohol timing, bedroom temperature, total time in bed, then training load. Each of those has more leverage than any supplement, and each shows up in the data within a week.
Also watch for orthosomnia — anxiety generated by chasing the score itself. People who check the number before they are fully awake report worse subjective sleep on days the app shows something they dislike. Reading the number an hour after waking, weekly rather than daily, removes most of that effect.
When to talk to a physician
Take the numbers to a clinician if you sleep 7–8 hours, show 65%+ core with deep under 10% for several weeks, and still wake unrefreshed. That combination with loud snoring, witnessed breathing pauses, morning headaches, dry mouth on waking, or overnight blood oxygen dipping below 90% is a reasonable prompt for a home sleep apnea test. An estimated 80% of moderate-to-severe obstructive sleep apnea is undiagnosed, and the most-missed presentation is the lean, fit adult with a narrow airway — apnea is not exclusively a weight condition.
Two medication interactions are worth raising with a prescriber rather than solving alone. SSRIs and SNRIs suppress REM at ordinary doses, which mechanically raises the core share on your watch. Z-drugs such as zolpidem increase N2 while doing little for slow-wave sleep, so a prescription sleep aid can push core up and deep down at the same time. Neither is a reason to stop a medication; both are reasons not to interpret the chart as a lifestyle failure.
Frequently Asked Questions
What is core sleep on Apple Watch?
Core sleep is Apple's label for combined N1 and N2 sleep — the light-to-moderate stages that make up the largest share of the night. Expect roughly 50–60% of total sleep time, or about 3.5 to 4.5 hours on a 7.5-hour night. Apple chose the word "Core" because these stages form the structural bulk of every 90-minute cycle. The underlying classification is the same four-bucket model every other major tracker uses.
Is core sleep the same as light sleep?
Yes, functionally. Apple Watch reports "Core" where Fitbit, Garmin, Oura, and Whoop all report "Light" — listed alphabetically, and all describing the same N1 + N2 stages. The numbers still will not match across devices, because each brand uses different sensors and a different classification model. Comparing your Core percentage on one device against a friend's Light percentage on another tells you almost nothing.
Is a lot of core sleep bad?
Not by itself. Core is supposed to be the majority of the night, so a high core number usually reflects arithmetic rather than a problem. The exception is when core rises because deep and REM both fell — for example, core at 75% with deep under 8%. That pattern shows up reliably after alcohol, in a bedroom above 72°F, and in untreated sleep apnea, and it is worth investigating.
Why does my core sleep look higher on some nights than others?
Total sleep time is the biggest driver. Cut a night short and you lose REM disproportionately, which pushes core's share up even though core minutes fell. Sensor fit matters too: a loose watch band reads more movement, and movement pushes the model toward light-stage classification. Wearing the watch snugly one finger-width above the wrist bone reduces that artifact.
Can I increase deep sleep at the expense of core sleep?
You cannot choose the split directly, but you can shift it. The changes with real leverage are a fixed wake time seven days a week, no alcohol within four hours of bed, and a bedroom at 65–68°F. Those three typically move deep sleep 10–25 minutes over a couple of weeks, which lowers core's share as a side effect. Anything sold as a "deep sleep booster" moves the number far less than room temperature does.
Should I trust the core sleep number at all?
Trust the direction, not the digits. Validation work generally puts consumer stage classification at 60–80% agreement with polysomnography, and stage boundaries are exactly where the models are weakest. Total sleep time on the same devices is usually accurate to within 10–20 minutes, which is why duration and consistency deserve your attention before any stage percentage does.
Related
- Sleep guides — stages, scores, and what moves them
- How much deep sleep do I need?
- REM sleep explained
- Sleep score explained
- Apple Watch sleep score
- Apple Watch vs Oura