Wearable studies: what the data can and cannot show

Two wearable studies, two very different claims. The first came from an interdisciplinary team at The University of Texas at Austin and appeared in Nature Scientific Reports. The team tracked 82 young adults over several months. A wrist-worn activity tracker recorded movement and heart rate. A smartphone app collected self-reported well-being.

The finding, according to UT News: physical activity lengthened REM latency, the time it takes to enter REM sleep. The researchers argue this may mean exercise helps consolidate deeper sleep stages before the brain transitions into REM, the stage where dreaming is vivid and brain activity resembles waking.

That is a modest, mechanistic claim. It is not the same as saying a wearable can tell you whether you slept well.

What the UT study actually did

The design matters because the researchers say it addresses a known weakness. Most sleep studies run in labs and draw conclusions from a single night. Benjamin Baird, a research assistant professor of psychology and one of the authors, said lab settings are unfamiliar and stressful, and that "you can't really look over time." The team used the wearables to monitor participants at home, at work and during daily activities, for weeks or months.

David M. Schnyer, co-author and chair of the Department of Psychology, framed the device as the point. "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, and in a way that's showing predictive results," he said. Researchers can now study sleep architecture in the field rather than the lab, he added.

The study replicated earlier lab findings: low-intensity and moderate-to-vigorous activity were linked to deeper, more restorative sleep, and better sleep was associated with more energy and less stress the following morning. Those are associations from 82 people wearing consumer hardware.

The stress study and its rebuttal

A second study got a rougher reception. Wareable reported on 12 August 2025 that researchers tracked stress, fatigue and sleep for three months in 800 young adults wearing Garmin watches. Participants checked in four times a day on how stressed, fatigued or sleepy they felt. The study found "basically zero" correlation between the device's stress score and self-reported feelings. For a quarter of participants, the device said the opposite of what they reported.

The Guardian summarised it as smartwatches that cannot accurately measure stress. Other outlets went further, claiming watches "confuse stress with exercise."

Wareable's objection was partly about hardware age. The benchmark was the Garmin Vivosmart 4, released in 2018, and Wareable called that "an absurd" choice, comparing it to testing a 2018 Kia and drawing conclusions about electric vehicles in 2025. The publication also argued that exercise is itself acute physiological stress, so a device flagging a workout as high-stress is working, not confused.

"The single-night tests are hugely noisy and inaccurate," said Danny Eckert, director of the Adelaide Institute for Sleep Health at Flinders University.

That quote comes from ABC News, which reported on 24 September 2024 that in-home monitoring has opened the door to large-scale sleep research. A 2022 study co-authored by Eckert found that up to half of single-night diagnostic studies for sleep apnoea, the current gold standard, misclassify severity.

ABC also noted that Samsung's smart watch sleep apnoea detection system received US regulatory approval in early 2024 after an eight-month review, the first consumer smart watch to clear that hurdle. Apple's watch received the same approval later that month.

The gap between signal and interpretation

Put the two studies side by side and the pattern is not that wearables fail. It is that the claims made about them often outrun the evidence. UT researchers say a consumer Fitbit is sensitive to sleep architecture measures in a field setting. The stress study says one algorithm on a seven-year-old tracker did not match how 800 people felt. Both can be true.

Wareable's own conclusion was that a device flagging a real biological event may be right while the wearer's interpretation lags. That is an opinion, not a finding. But it lands close to what the UT team is arguing: the value of wearable data depends on what question you ask of it, and on whether the study was designed to answer that question.