Sleep efficiency variability
+1.77%
Greater within-subject SD
Individuals with chronic insomnia showed 1.77% greater night-to-night variability in sleep efficiency than good-sleeper controls (p<.001).
Insomnia variability study · 2026
Variability defines chronic insomnia.
A new longitudinal study from Sleep.ai, Washington State University, and the University of Washington followed 112 adults for 8 consecutive weeks of contactless monitoring. The defining marker of chronic insomnia was how unpredictably people slept from one night to the next.
- 112 adults
- 8 weeks at home
- Peer-reviewed
- JMIR · 2026
Nightly sleep efficiency, 8 weeks
Individuals with chronic insomnia
Good-sleeper controls
Key takeaway
Both groups slept about the same total hours each night. The defining marker of chronic insomnia is night-to-night unpredictability.
Hansen et al., JMIR Form Res 2026 · doi:10.2196/73969Three findings
What 8 weeks of nightly contactless monitoring revealed.
Adults with chronic insomnia and good-sleeper controls were tracked across 56 consecutive nights in their own homes. Across three core sleep measures, individuals with chronic insomnia showed significantly greater night-to-night variability at p<.001 (a result this strong would happen by chance less than 1 in 1,000 times).
Sleep latency variability
+8.80
Minutes greater within-subject SD
Individuals with chronic insomnia showed 8.80 minutes greater night-to-night variability in time to fall asleep (p<.001).
Intermittent wakefulness variability
+8.60
Minutes greater within-subject SD
Individuals with chronic insomnia showed 8.60 minutes greater night-to-night variability in time spent awake during the night (p<.001).
Implications
What this means for sleep health products.
Variability is a clinically meaningful endpoint, and most sleep products in market today do not measure it. The implications run across consumer sleep tech, sleep medicine, and brands developing validated sleep claims.
Consumer sleep tech
Variability is the missing variable in your user data.
Apps in women’s health, fitness, nutrition, and mental health live or die on engagement, and users disengage when their progress feels random. A user whose sleep oscillates from 4 to 9 hours night-to-night is the same user whose hormone tracking, training load, GLP-1 response, or mood data look chaotic. Surfacing variability alongside duration gives users a measure that matches what they’re feeling, and gives your product a richer signal to act on.
Sleep medicine
Sleep.ai’s contactless tech captures what one-night studies miss.
Polysomnography over a single night cannot detect a pattern that only emerges over weeks. The Insomnia Severity Index relies on memory and averages. Sleep.ai’s contactless radiofrequency sensing technology, built on twelve years of patented ResMed research and PSG-validated, captures objective sleep night after night in patients’ actual homes. No wearable, no electrodes, no observer effect. It fills a measurement gap in chronic insomnia that has been open for decades.
Brand R&D
Variability reduction may be your most defensible claim.
Brands developing supplements, beverages, mattresses, devices, and wellness platforms often run validation studies in lab settings or over windows too short to capture what their product does in people’s actual lives. Studying sleep and insomnia at home, over weeks, with contactless monitoring lets R&D teams detect within-subject variability reduction. That endpoint is often the strongest signal a product produces and the most defensible claim a brand can take to market.
Methodology
How the study was designed and run.
A prospective cohort study comparing individuals with chronic insomnia to good-sleeper controls across 56 consecutive nights of nightly, at-home, contactless sleep recording. Published in JMIR Formative Research on March 18, 2026.
- Participants 112 adults (83 with chronic insomnia, 29 healthy good-sleeper controls)
- Duration 8 consecutive weeks of nightly recording in participants’ homes
- Method Contactless ultra-low-energy radar, no wearables, naturalistic home environment
- Instrument SleepScore Max, validated against polysomnography
- Statistical analysis Mixed-effects models controlling for age and sex
- Trial registration NCT04013321
- Funding NIH grant KL2TR002317
- Publication JMIR Formative Research, March 18, 2026 · DOI 10.2196/73969
The researchers
A multi-institution research team led by sleep medicine experts.
The study was led by Devon A. Hansen, PhD, of Washington State University’s Sleep and Performance Research Center, and Nathaniel F. Watson, MD, MSc, of UW Medicine Sleep Center at the University of Washington. Sleep.ai authors Elie Gottlieb, PhD, Sharon Danoff-Burg, PhD, and Roy Raymann, PhD contributed to study design and execution. Additional contributors include Mary E. Peterson, MA (Claremont Graduate University), Myles G. Finlay, BS (Washington State University), and Dedra Buchwald, MD (University of Washington).
“People with chronic insomnia live with unpredictable sleep night after night. Being able to track this objectively in people’s own homes over two months opens up new possibilities for both research and care.”
Sleep and Performance Research Center
“This study validates that meaningful sleep insights require more than a single night’s snapshot. By tracking sleep objectively and contactlessly in people’s own homes, we can move beyond lab-based limits and give consumers and clinicians tools to understand sleep as it truly happens, night after night.”
Study co-author
Get the full study
Read the peer-reviewed paper in full.
The full 2026 JMIR Formative Research publication. Includes the complete methodology, all statistical results, and the discussion of what these findings change for sleep tech, sleep medicine, and brand R&D.
- The exact methodology Sleep.ai used to measure sleep at home, over 8 weeks, with no wearable
- Every statistical comparison between individuals with chronic insomnia and good-sleeper controls
- A peer-reviewed reference you can cite in product positioning, claims, and regulatory work
- Discussion of what variability changes for sleep tracking products and clinical research
Send me the JMIR study
Loading form…
Frequently asked
Questions researchers and product teams ask about this study.
What is the main finding of the JMIR insomnia variability study?
Individuals with chronic insomnia experience significantly greater night-to-night variability in sleep efficiency, sleep latency, and intermittent wakefulness than good-sleeper controls. Mean sleep duration was similar between groups (6.59 vs. 6.60 hours), so unpredictability is the defining marker of the chronic insomnia phenotype. The finding is published in JMIR Formative Research (Hansen et al., 2026, DOI 10.2196/73969).
How is night-to-night sleep variability different from average sleep duration?
Average sleep duration measures how much someone sleeps over a given period. Night-to-night variability measures how much that amount swings from one night to the next. Two people can average the same 6.5 hours of sleep, but one might consistently sleep 6 to 7 hours while the other oscillates between 4 and 9. The Hansen et al. study found this swing, captured as within-subject standard deviation across 8 weeks, distinguishes chronic insomnia from healthy sleep.
How was sleep measured in the study without wearables?
Sleep was measured using the SleepScore Max, a contactless device that uses ultra-low-energy radio-frequency sensing to detect breathing, body movement, and sleep parameters from a bedside table. The instrument requires no contact with the body, no app worn at night, and no electrodes. It has been validated against polysomnography (Chinoy et al., 2021, SLEEP) and against wrist actigraphy in individuals with chronic insomnia (Teeter et al., 2022).
Who led the longitudinal chronic insomnia variability study?
The study was led by Devon A. Hansen, PhD, of Washington State University’s Sleep and Performance Research Center, and Nathaniel F. Watson, MD, MSc, of UW Medicine Sleep Center at the University of Washington. Sleep.ai scientists Elie Gottlieb, PhD, Sharon Danoff-Burg, PhD, and Roy Raymann, PhD, contributed to study design and execution. The work was funded by NIH grant KL2TR002317 and registered as NCT04013321.
Why does an 8-week monitoring period matter for studying insomnia?
Shorter studies miss the variability signal entirely. Earlier work using one to three weeks of polysomnography or actigraphy struggled to detect the night-to-night swings that define chronic insomnia. Wohlgemuth et al. observed that up to three weeks of recordings were needed to demonstrate stability in measures like wake after sleep onset and sleep latency. The 8 consecutive weeks of contactless recording in this study allowed within-subject standard deviations to stabilize, making the variability signal visible at high statistical confidence (all p<.001).
What does this study mean for sleep tracking products and clinical research?
Variability is a clinically meaningful endpoint, and most sleep products on the market today do not measure it. Apps that surface only nightly averages or 7-day rolling means hide the volatility that users with chronic poor sleep experience. Validation studies for supplements, devices, and digital health programs that track only group-level mean changes may miss the within-subject variability reductions that most accurately reflect a product’s clinical effect. Long-term, naturalistic, contactless measurement opens a more rigorous endpoint for sleep tracking products and the validation work behind sleep claims.
What kinds of sleep research can be run with this contactless method?
Sleep.ai partners with consumer apps, supplement brands, mattress companies, fitness platforms, and digital health programs to design and run validated sleep research at scale. The Hansen et al. study established that the SleepScore Max instrument can detect both group-level mean differences and within-subject variability differences in chronic insomnia at p<.001, making it suitable for product validation studies, longitudinal outcome research, and any work where naturalistic, at-home, multi-week sleep measurement is more meaningful than a single night in a lab. Teams evaluating sleep endpoints can contact Sleep.ai’s research team to discuss study design.
Related research
Further reading on contactless sleep measurement and chronic insomnia.
Build with Sleep.ai
Build your product on validated sleep science.
Sleep.ai works with consumer apps, supplement brands, mattress companies, fitness platforms, and digital health programs in two ways: embedding validated sleep science into your product by plugging into our sleep intelligence platform and designing the kind of published research that turns your product into a defensible, citable sleep claim. Studies run in participants’ actual homes, over weeks, and reach peer-reviewed publication.
~1B
Hours of real-world sleep data
250+
Scientific studies
100+
Peer-reviewed publications
74M
Covered lives via Dein Schlaf in Germany
Citation Hansen DA, Peterson ME, Finlay MG, Gottlieb E, Danoff-Burg S, Raymann RJEM, Buchwald D, Watson NF. Assessing Night-to-Night Sleep Variability as a Hallmark of Chronic Insomnia Using Longitudinal, Contactless, Mobile Sleep Monitoring: Prospective Cohort Study. JMIR Form Res 2026;10:e73969. doi: 10.2196/73969