Wearable devices have made it possible to see data about your sleep that previously required a polysomnography lab. Apple Watch, Fitbit, Garmin, and Oura all estimate sleep stages, track heart rate through the night, and produce a daily sleep score. For perimenopausal women, who often know something is wrong with their sleep long before they have language for it, this data can be useful.

It can also be confusing, and sometimes anxiety-inducing. This article explains what consumer wearables actually measure, where they are accurate, where they are not, and how to interpret the data in the specific context of perimenopausal sleep.

How Wearables Estimate Sleep

Consumer wearables cannot measure brain activity. Clinical sleep studies (polysomnography, PSG) classify sleep stages using electroencephalography (EEG), electromyography (EMG), and electro-oculography (EOG). Wearables use a different set of inputs: movement (accelerometer), heart rate (photoplethysmography, PPG), and sometimes skin temperature and oxygen saturation (SpO2).

From these signals, algorithms infer sleep stages. The inference is probabilistic, not direct. During light sleep (N1 and N2), heart rate variability tends to be lower than during deep sleep (N3, slow-wave sleep), and movement tends to be minimal but slightly more frequent than during deep sleep. During REM, movement is low (REM atonia temporarily paralyses the large muscle groups), but heart rate is more variable and irregular.

Independent validation studies comparing wearable sleep staging to simultaneous PSG have found that consumer devices perform well at distinguishing sleep from wakefulness (typically 85–90% accuracy), but are less reliable at staging individual sleep phases.[1] Deep sleep (N3) tends to be underestimated by most devices; REM tends to be approximately correct; light sleep is overestimated as a consequence.

A 2019 study comparing Apple Watch, Fitbit, and Oura to PSG found that all three devices identified sleep onset reasonably well and were moderately accurate for total sleep time and REM percentage, but all showed significant variability on individual nights and across participants.[2]

The practical implication: do not treat any single night's staging breakdown as precise. Look at trends over time.

What Perimenopausal Sleep Changes Wearables Can Detect

Several changes in sleep architecture that occur in perimenopause produce signals wearables can detect.

Increased wake episodes. Vasomotor events (hot flashes, night sweats) that wake you briefly will register as awakenings in your data. More fragmented sleep scores without any change in bedtime habits or alcohol intake often reflect a vasomotor signature.

Reduced deep sleep percentage. Deep sleep (slow-wave sleep) declines with age in both sexes, but the decline accelerates in perimenopause. Estrogen and progesterone both support deep sleep; as both decline, slow-wave activity decreases. Wearables already underreport deep sleep, so if your device is showing low deep sleep percentages, the true deficit may be smaller, but the trend direction is informative.

Heart rate variability overnight. Wearables that track HRV (Oura, Garmin, Apple Watch) measure HRV during the early part of sleep when it tends to be most stable. Vasomotor events cause brief increases in sympathetic activity and corresponding dips in HRV, which may show up as lower overnight HRV on nights with significant hot flashes.

Elevated resting heart rate overnight. On nights with frequent hot flashes, resting overnight heart rate tends to be slightly elevated compared to your baseline, reflecting the cardiovascular component of vasomotor events.

Skin temperature deviation. Devices with skin temperature tracking (Oura, Fitbit Sense, Google Pixel Watch) measure nightly skin temperature variation. Hot flashes produce a brief spike followed by a decline as sweating begins. Consistent elevated or erratic overnight skin temperature readings can reflect vasomotor activity, though the signal is noisy and device sensitivity varies.

How to Use the Data

Single-night scores carry too much noise to be meaningful on their own. A week-over-week trend in sleep efficiency (time asleep as a percentage of time in bed), average overnight wake duration, and resting heart rate is far more informative. Most platforms allow you to view 30-day or 90-day trends; use these rather than fixating on the previous night.

Correlate with symptoms

The most useful thing you can do with wearable sleep data in perimenopause is correlate it with your symptom log. If your data shows you logged several hot flashes after 2 AM, and your wearable shows fragmented sleep and elevated heart rate in the same window, that is meaningful confirmation. If your sleep score is poor but you have no obvious symptoms, that is a different puzzle; sleep apnea, which worsens at menopause, is worth considering (see below).

Set your own baseline

Published reference ranges for sleep stages are population averages. They include people of all ages, both sexes, and varying health status. Your personal baseline across weeks of stable data is more relevant than any published norm. A meaningful change from your baseline is more actionable than comparing yourself to a population average.

Watch for orthosomnia

Orthosomnia describes insomnia triggered or worsened by obsessive monitoring of sleep data.[3] Sleep tracking can worsen sleep quality in anxious individuals by increasing pre-sleep monitoring behaviour and post-wake data checking. If you find yourself checking your device immediately on waking and feeling worse depending on the number you see, consider checking only weekly rather than nightly, or turning off the summary notification.

The goal is information that supports better decisions, not a new source of performance anxiety about something you cannot consciously control.

Sleep Apnea at Menopause

Obstructive sleep apnea (OSA) is significantly underdiagnosed in women, partly because women's symptoms differ from the classic male presentation (loud snoring, observed apnea). Women tend toward non-restorative sleep, fatigue, mood disturbance, and insomnia rather than the textbook picture.

OSA prevalence increases substantially after menopause. Data from the Wisconsin Sleep Cohort Study found that postmenopausal women not using HRT had a 3.5-fold greater odds of moderate-to-severe OSA compared to premenopausal women, with HRT use associated with lower OSA risk.[4]

Estrogen and progesterone both support upper airway muscle tone. As they decline, pharyngeal muscles relax further during sleep, increasing the likelihood of airway narrowing or collapse.

Wearables with SpO2 monitoring (Apple Watch Series 6 and later, Fitbit Charge 5 and later, Oura Gen 3 and later) can detect oxygen desaturations during sleep, a downstream consequence of apneic events. A device that consistently shows SpO2 dropping below 90%, or frequent large drops from your typical overnight baseline, warrants a conversation with your GP about formal sleep apnea assessment.

Consumer SpO2 tracking is not a diagnostic tool for OSA, but it can identify a pattern worth investigating. Many NHS areas offer community sleep studies that are far more accessible than specialist sleep clinics.

What to Bring to a Clinical Appointment

If you are discussing sleep problems with a GP or menopause specialist, wearable data can supplement subjective reporting. Consider bringing:

  • A 2–4 week export or screenshot of sleep efficiency and wake trends
  • Any correlations you have noticed between symptom data and sleep data
  • Your average overnight heart rate trend compared to a baseline period

Frame this as supplementary information, not a diagnosis. "My wearable has been showing more frequent wake episodes and lower HRV for the past month, which matches my subjective experience" is useful clinical context. A clinician may not be able to interpret every number on your report, but trend data presented clearly is informative.

References

[1] Depner, C. M., Cheng, P. C., Devine, J. K., et al. (2020). Wearable technologies for developing sleep and circadian biomarkers: A summary of workshop discussions. Sleep, 43(2), zsz254. https://doi.org/10.1093/sleep/zsz254

[2] de Zambotti, M., Rosas, L., Colrain, I. M., Baker, F. C. (2019). The sleep of the ring: Comparison of the Oura sleep tracker against polysomnography. Behavioral Sleep Medicine, 17(2), 124–136. https://doi.org/10.1080/15402002.2017.1300587

[3] Baron, K. G., Abbott, S., Jao, N., Manalo, N., Mullen, R. (2017). Orthosomnia: Are some patients taking the quantified self too far? Journal of Clinical Sleep Medicine, 13(2), 351–354. https://doi.org/10.5664/jcsm.6472

[4] Young, T., Finn, L., Austin, D., Peterson, A. (2003). Menopausal status and sleep-disordered breathing in the Wisconsin Sleep Cohort Study. American Journal of Respiratory and Critical Care Medicine, 167(9), 1181–1185. https://doi.org/10.1164/rccm.200209-1055OC