Why sleep tracker readings are estimates, not brain measurements

A review of 24 studies found differences from clinical sleep tests. Separate feedback trials help explain why feeling better about sleep is not the same as sleeping better.

A review of 24 studies found differences from clinical sleep tests. Separate feedback trials help explain why feeling better about sleep is not the same as sleeping better.
Wrist sleep trackers estimate sleep from bodily signals rather than record brain activity. They can help reveal general sleep patterns, but their readings are not equivalent to a clinical sleep test—and improved ratings of sleep do not necessarily mean longer or less interrupted sleep.
A meta-analysis—a study that combines results from several studies—published on March 1, 2025 in the Journal of Clinical Sleep Medicine compared wrist trackers with polysomnography across 24 studies involving 798 participants. Polysomnography records brain waves, eye movements and muscle tone using an electroencephalogram, or EEG, and other sensors.
The review found a pooled difference of about 17 minutes in total sleep time. Writing in The Conversation, Carnegie Mellon University psychology and neuroscience professor J. David Creswell describes this as trackers overestimating sleep. It is an average across the studies, not an error that applies equally to every device, person or night.
Creswell explains that trackers can mistake quiet wakefulness for sleep. Someone lying awake without moving may therefore be recorded as asleep. For people with insomnia, prolonged stretches of motionless wakefulness can inflate the estimated amount of sleep.
Sleep-stage readings are also indirect. According to Creswell, wearables use heart rate, changes in the time between heartbeats, movement and skin temperature. Their algorithms use patterns in those signals to infer a sleep stage; they do not measure the brain activity that defines it. He reports that a separate 2025 study of six wrist devices worn by 62 adults found only modest agreement with polysomnography on sleep stages.
Creswell disclosed in his October 2026 article that he served as chief of science at Equa Health, a mindfulness meditation mental-health app.
Accuracy is one question; what happens when people receive sleep feedback is another. The Conversation describes two trials in which guidance accompanied the reports. The outcomes distinguish subjective measures—how people rate their sleep or symptoms—from measurements of sleep itself.
| Trial | Comparison | Reported result |
|---|---|---|
| 2023 randomized trial: 113 adults with insomnia symptoms, followed for five weeks. | Wearable feedback plus help interpreting it, versus standard sleep-hygiene education: advice on habits such as regular bedtimes and reducing caffeine. | The feedback group reported milder insomnia symptoms, but did not sleep longer, fall asleep faster or wake less. |
| 2026 trial: 229 older adults in Japan, using monitors placed under their bedding rather than wrist trackers. | Sleep reports alone, versus reports plus monthly telephone guidance on sleep habits and daytime activity. | Reports alone did not improve insomnia scores. With guidance, participants’ ratings improved at least over the first three months. |
In the 2023 trial, feedback also did not reduce the gap between participants’ estimates of their sleep and the sensor readings any more than sleep-hygiene education did. The symptom improvement mattered to participants, but it was not a measured increase in sleep.
The meta-analysis found differences between device groups. For Fitbit, it found no statistically significant difference from polysomnography in time spent awake after first falling asleep. For the other-device group, it found no statistically significant difference in the time taken to fall asleep. These findings do not establish that either group measured those periods exactly.
The review searched for studies through March 2024 and referenced older evaluations published in 2015 and 2018. Creswell notes that tested devices have often been replaced by the time a study appears. The pooled findings therefore cannot be assumed to describe current models.
The review’s authors conclude that wrist trackers can still be useful for following general sleep patterns, while remaining less reliable than polysomnography for key measurements. They call for further improvements and clinical studies to establish better accuracy.

The ministry consultant also described free screening for women aged 45–68 through ArmMed, though the oncology centre’s guidance sets narrower eligibility and a different booking route.
3 min

Planning and soil surveys can continue, but major site changes must pause for environmental review. Available electricity capacity does not remove that requirement.

Distances between Earth-orbiting satellites cannot fully establish their positions. Simulated links to lunar orbiters gave researchers another way to constrain the error.
4 min
2 min