Wearable AI Tool Estimates Vascular Age for Cardiovascular Risk

Measuring it usually requires clinic-based equipment and trained staff, limiting access and frequency. This gap constrains early risk stratification and longitudinal surveillance. A new study shows that overnight pulse signals recorded by a consumer wearable ring can yield vascular age estimates during routine sleep, offering a passive and scalable approach to cardiovascular assessment.

Researchers at the Centre for Sleep and Cognition at the NUS Yong Loo Lin School of Medicine developed and validated an analytical pipeline to derive vascular age from sleep recordings captured by a finger-worn device. The team analyzed photoplethysmography (PPG) signals, the light-based pulse waveform used by fitness trackers to measure heart rate. Their goal was to transform passive sleep data into clinically interpretable cardiovascular insights.

The method processed PPG collected overnight by the Oura Ring, a consumer sleep tracker, and compared outputs with signals from a clinical fingertip sensor. Investigators evaluated both traditional feature-based models and a deep learning model to estimate vascular age. They built and validated their own pipeline independent of any proprietary device algorithms to ensure transparency and reproducibility.

The deep learning model predicted vascular age from both the wearable and the clinical fingertip sensor with similar accuracy. Reported mean error was six to seven years with strong agreement to participants’ actual ages. Ring-derived vascular age estimates were also associated with blood pressure, a standard cardiovascular health marker.

Findings were published in PLOS Digital Health on March 30, 2026. The authors noted that consumer wearables already in daily use could support scalable, longitudinal monitoring outside clinical settings. Future work will test performance across more diverse populations and assess whether wearable-derived vascular age can inform preventive care and clinical decision-making.

“Signals collected passively during sleep can be translated into clinically meaningful insights about vascular health. This opens the door to scalable, longitudinal monitoring of cardiovascular health using devices people already wear in their daily lives,” said Dr. Gizem Yilmaz, a research fellow and co-first author of the study at the Centre for Sleep and Cognition, NUS Medicine.

“Our findings lend credence to moving cardiovascular monitoring out of the clinic and into everyday life. Wearable-derived vascular age could, in time, support earlier detection of cardiovascular risk, reinforce positive lifestyle habits, and feed into large-scale population health studies,” said Professor Michael Chee, Director at the Centre for Sleep and Cognition, NUS Medicine, and the study's principal investigator.

Related Links
NUS Medicine