Wearables Turn the World Into a Sleep Lab, but the Evidence Is Uneven

The study appeared in Nature Scientific Reports, and UT Austin announced it on 2 April 2024. The team tracked sleep and activity in 82 young adults over several months. Participants wore a wrist tracker that logged movement and heart rate. A separate smartphone app collected self-reported well-being data. The headline result: physical activity lengthened REM latency, the time it takes to enter REM sleep. The researchers read that as a sign exercise may help consolidate deeper sleep stages before the brain shifts into REM.

That is a small sample. It is also a specific one: young adults, tracked at home rather than in a clinic.

The methodology is the real story. Sleep research has leaned on lab studies that observe a single night, an approach the UT team describes as a known limitation. "You can learn a lot from lab studies, but obviously there are limitations to studying the sleep patterns of individual participants in just one night," said Benjamin Baird, a research assistant professor of psychology and one of the authors. The unfamiliar, clinical-type setting can be stressful, he added. A single night cannot show change over time, he said, which leaves open questions about generalizability.

Sleep architecture, the structure of each 90 to 120 minute cycle, includes three stages of non-REM sleep (light, deep and deepest NREM) plus REM, which makes up roughly the final 25% of a cycle, according to the UT release. The team says its wearable setup could detect these stages well enough to produce predictive results. "We've shown using a standard Fitbit that anyone could wear, not even an expensive scientific device, that it is actually sensitive to these sorts of sleep architecture measures," said David M. Schnyer, a co-author and chair of the Department of Psychology.

The study replicated several lab findings in the wild. Both low-intensity and moderate-to-vigorous physical activity were linked to deeper, more restorative sleep. Better sleep, in turn, went with more energy and less stress the following morning. The pilot work came out of Whole Communities-Whole Health, a program at UT Austin that takes an interdisciplinary approach to health data collection.

That pattern, consumer devices feeding research-grade questions, is now global. The ABC reported on 24 September 2024 that when a powerful earthquake hit New South Wales, data scientists in Sweden detected an above-average number of people briefly waking in the surrounding area. They based that on data from the Swedish sleep-tracking app Sleep Cycle. Michael Gradisar, the company's head of science, told the ABC that sleep is one of the first things to go wrong in a person's health. He described it as the canary in the mine.

Sleep Cycle has about 2 million active nightly users. Gradisar joined from academia to work on that data. During the Omicron wave in New York, the ABC reported, the company's data scientists saw an increase in night-time coughing roughly one week before confirmed cases rose. He described that as an early warning system.

Regulators have started to move with the hardware. In early 2024, Samsung's smartwatch sleep apnoea detection system received approval from the US health regulator after an eight-month review, the first consumer smartwatch to clear that hurdle. Apple's smartwatch received the same approval later that month, the ABC reported. That matters because sleep apnoea may affect as many as a billion people worldwide and is under-diagnosed, according to the ABC's reporting. The conventional diagnosis requires an inconvenient, often expensive night in a sleep lab.

Not everyone treats the shift as a straight upgrade. Gabriel Pires, a sleep researcher in Brazil, told the ABC that he and colleagues started publishing papers warning people to be careful with the new technologies because they did not work well. Gradisar was also sceptical before moving to the private sector. The accuracy of some devices improved after the pandemic, when sales of smartwatches grew and investment followed. Apple added sensors for heart electrical activity, blood oxygen and body temperature.

Even the gold standard has problems. A 2022 study co-authored by Danny Eckert, director of the Adelaide Institute for Sleep Health at Flinders University, found that up to half of single-night diagnostic studies for sleep apnoea misclassify severity. "The single-night tests are hugely noisy and inaccurate," he told the ABC. In-home monitoring now makes it possible to track sleep apnoea every night for weeks, months and years, he added.

The UT study makes a narrower claim than the consumer-tech boom around it. It does not say wearables can diagnose anything. It says a standard tracker was sensitive to sleep architecture measures in a real-world setting, in 82 people, and that the resulting data lined up with prior lab findings. The authors frame that as an opening. "The world is your oyster now," Schnyer said. "You can use this device to study all manner of different sleep architecture data related to lifestyle, related to mood and mood disorders, in the field, not in a lab, that people might have thought was not possible previously."

Two cautions are worth keeping in view. First, self-reported well-being data from a smartphone app is not the same as a clinical measure. The study's sample is young and small, so the effect sizes should not be read as population-level. Second, the accuracy debate around consumer wearables has not been settled by one approval or one paper. Pires's warnings and Eckert's critique of single-night testing point in the same direction: more nights, more people, and clearer reporting about what the devices can and cannot see.

What has changed is the scale of the instrument. Less than a decade after the first sleep-tracking Apple Watch, the world is wired for sleep study, as the ABC put it. David Walsh, who built the first Apple Watch sleep-tracking app, told the ABC that testing his own creation made him acutely aware of how behaviour affected sleep, down to an after-work beer. That kind of continuous, longitudinal data is what the UT team used to argue that sleep architecture can be studied outside the lab. Whether it produces better medicine depends on studies that are larger, longer and more sceptical than the ones that came before.