What core sleep means on a wearable device

When you check your wearable device in the morning, core sleep often occupies the largest block on the screen. Many people wonder whether this label refers to deep rest or an essential stage that requires a specific nightly total. In consumer technology, the term is simply a proprietary name for light sleep rather than a newly discovered biological phase.

What core sleep represents on consumer trackers

In a 2024 single-night inpatient study that compared three consumer sleep trackers with polysomnography, the researchers sorted every device's output into wake, light sleep, deep sleep and REM, defined light sleep as polysomnography stages N1 and N2, and counted the Apple Watch's "Core" category as light sleep. ^[1]

A 2023 study of 11 consumer sleep trackers treated the label the same way: when it standardised every device's stage codes into wake, light, deep and REM before comparing them with polysomnography, it noted that the Apple Watch 8 used the expression "core sleep" instead of light sleep. ^[2]

The historical research definition of core sleep

In sleep research the term has been used with a different meaning. A 1992 study of 24-hour recordings in 14 healthy people aged 88 to 102 and 19 young adults aged 25 to 35 used a distinction proposed by Horne between core and optional sleep, in which core sleep is the essential part of sleep, mainly slow wave sleep made up of non-REM stages 3 and 4, obtained during the first three sleep cycles, and the rest of the night is optional sleep. ^[3]

CIBDOL · The historical research definition of core sleep

The stage names changed with the scoring rules. A 2009 study scored the all-night recordings of 72 healthy people and patients aged 21 to 86 twice, under the 1968 rules and under the 2007 rules, which it reports as stage 1 becoming N1, stage 2 becoming N2 and slow wave sleep, stages 3 and 4 together, becoming N3. Time spent in each of these stages differed significantly between the two rule sets, while sleep latency, REM latency, total sleep time and sleep efficiency did not. ^[4]

Tracker measurements compared with polysomnography

In the 2024 study, 35 healthy adults aged 20 to 50 spent one night in an inpatient study wearing all three devices, and Apple Watch data were analysed for 29 of them. Compared with polysomnography, the Apple Watch Series 8 overestimated light sleep by 45 minutes and underestimated deep sleep by 43 minutes. Epoch by epoch, it identified on average 86.1 percent of polysomnography's light sleep epochs, and on average 72.7 percent of the epochs it labelled light sleep were light sleep on polysomnography. ^[1]

In the same study, agreement between the three devices and polysomnography on each night's total of light sleep was poor to fair, with intraclass correlations of 0.37 to 0.52, and it was poor for the nightly totals of deep sleep and REM sleep. ^[1]

A living systematic review and meta-analysis of 82 studies that validated Apple Watch measurements against a criterion method, with searches run to 24 September 2025, judged accuracy for sleep, as for step count, to be moderate, and found that accuracy varied by metric, measurement conditions and individual physiology. ^[7]

Device tendencies toward the light stage

The 11-tracker study, which enrolled 75 participants at a hospital and a sleep clinic in Korea, also reported where the errors went. Averaged across all devices, the predictions were biased toward the light stage, and the wearables among them mainly misclassified wake as light sleep. ^[2]

In a 2025 study in which 62 adults each wore two to four of six wrist-worn trackers during one night of polysomnography, most devices differed significantly from polysomnography in their estimates of light sleep, and the main epoch-by-epoch errors were wake, deep sleep and REM epochs that the devices classified as light sleep. ^[6]

Scoring variability and the absence of nightly targets

Human scorers also disagree most about the lightest stage. In a reliability programme in which more than 2,500 scorers made over 3.2 million scoring decisions on 1,800 epochs, agreement with the majority score was lowest for stage N1, at 63.0 percent, while agreement on N2 approached that for REM, the stage with the highest agreement. The first N2 epoch after N1 was among the epochs scorers found particularly difficult, and so was telling N2 from N3. ^[5]

This article explains what the label corresponds to and how well it is estimated. It sets no target for core or light sleep, does not say that any amount is good or bad, names no condition and recommends no device.