A team of researchers based out of University of Amherst, MA, USA say they’ve come up with an app that will supercharge sleep specialists’ ability to track our progress through different stages of shuteye. The app, developed by biomedical engineers, will allow academics to turn a regular Apple Watch into a sleep laboratory on your wrist. In an article published in IEEE Transactions on Biomedical Engineering, October 2025, scientists presented their novel approach to sleep monitoring.
The squad of research engineers, led by Professor Joyita Dutta, developed the app, dubbed BIDSleep, as part of a wider project to study how sleep and Alzheimer’s disease intersect. Routine sleep studies get expensive fast, with participants either needing to sleep in a laboratory full of specialized equipment or donning a bonnet of electrodes in their bed at home. Dutta needed to find a way to study multiple participants over an extended time – a challenge to both budget and bed space.
In a bid to make sleep monitoring a feasible endeavour, the researchers realized that they could create an app using existing tech. It uses data from the movement tracker and heart rate tracker already available in a consumer smart watch to follow us through each stage of sleep.
Dutta explains, ‘Our goal was to get as rugged as possible with a non-specialized consumer wearable device, which is the Apple Watch.’
The app, freely available to researchers, collects information about changes in our heart rate and movements throughout the night and uses a library of pre-recorded sleep data to figure out which stage of sleep a person is in. Using a huge pool of data compiled from over 800 subjects in a previous study, the team used AI to spot characteristic changes in motion and heart rate that matched up to three distinct stages of sleep.
Participants wore an Apple Watch to bed each evening. The researchers then down loaded information about activity levels and pulse using the regular accelerometer and heart rate tracker that the Apple Health suite of apps uses to track our fitness levels.
The researchers used data generated with specialized sleep monitoring equipment, including ECG, actigraphy and EEG, amongst other measures to find signatures that corresponded to each stage of sleep. Then they used AI driven statistical modelling to figure out how the data captured by the Apple watch lined up with them.
The BIDSleep app allows researchers to use an individual person’s Apple watch readings to identify when they were in each stage of sleep at any time of day or night. The app is still undergoing improvements, but the researchers say it is around 70% accurate. While it might not be ready for general use as consumer sleep tracker, given its continuing updates, Tzu-An Song, a postdoctoral research fellow on the team, is confident that it’s only a matter of time before the app is a staple in sleep studies. He says, ‘Overall accuracy matters, but sometimes we also need to look at the clinical metrics like sleep efficiency and sleep onset latency, total sleep time…our method works better for basically all of these metrics.’
Sleep quality has proven increasingly important as a way to predict our future health and the onset of all manner of conditions, including, as we reported this week, relapsing depression. It seems we might be seeing sleep study-approved wearable sleep trackers sooner than expected.
Song TA, Zhang Y, Zhou Z, et al. AI-Driven Sleep Staging Using Instantaneous Heart Rate and Accelerometry: Insights From an Apple Watch Study. IEEE Transactions on Biomedical Engineering. 2026;73(4):1596-1608. doi:10.1109/TBME.2025.3612158
Song TA, Chowdhury SR, Malekzadeh M, et al. AI-Driven sleep staging from actigraphy and heart rate. PLOS ONE. 2023;18(5):e0285703. doi:10.1371/journal.pone.0285703
An app, an Apple Watch and AI: UMass Amherst creates a new way for researchers to study sleep health. EurekAlert! Accessed April 17, 2026. https://www.eurekalert.org/news-releases/1103900
