AI models boost heart disease risk prediction from sleep data and sensors

A multidisciplinary research team has created an AI foundation model that analyzes routine sleep study data (polysomnograms) to predict long‑term health risks, including heart disease, cognitive decline and five‑year mortality. Published in Nature Communications on August 3 2026, the model identifies five distinct risk categories and outperforms the traditional apnea‑hypopnea index, with the highest‑risk group showing a 100 % increase in five‑year mortality risk.

Separately, an international team led by Ateneo de Manila University developed an AI system that estimates cardiac index from simple skin‑attached sensors. The model achieved 97.78 % accuracy in tests, offering a low‑cost alternative to conventional heart‑function assessments that require specialized equipment and trained staff. Led by Patricia Angela R. Abu, the technology could expand detailed heart monitoring to clinics in underserved regions.