A foundation model of wearable pulse oximetry reveals physiological signatures of health and cardiometabolic risk
- 1Weizmann Institute of Science
Abstract
While Photoplethysmography (PPG) is established as a noninvasive optical tool for monitoring heart rate and oxygen saturation, its high-resolution pulsatile blood volume waveforms contain rich physiological data that extend far beyond conventional vital signs.
We introduce PulseOx-FM, a foundation model, trained using self-supervised learning on 6,995,558 segments of pulse oximetry signals collected during 42,282 overnight sleep monitoring recordings of 10,704 adult participants in the Human Phenotype Project (HPP).
Using chronological age as a global health benchmark, PulseOx-FM significantly outperformed existing open-source and proprietary feature extraction methods while demonstrating robust generalization in an external out-of-distribution cohort. PulseOx-FM representations predicted 71 phenotypic targets spanning cardiometabolic, and neuropsychiatric domains beyond age, sex and body mass index. Applied in the same way to 3,721 intensive-care patients, two-minute windows of routine bedside pulse oximetry from the first day of admission improved the prediction of survival beyond age, sex and the heart rate and oxygen saturation the same device reports. Nightly embeddings further tracked next-day glycemic, dietary and activity-based state within individuals, dissociating this signal from sleep architecture alone. This next-day glycemic signal was predominantly a direct physiological effect, not explained by next-day dietary intake.
These findings suggest that PulseOx-FM provides a generalizable framework for encoding physiological patterns from sleep, offering a non-invasive tool for global health risk stratification and precision medicine.
Related
- Multi-Modal Wearable and Biological Data Integration Reveals Physiological Signatures of Weight Variation in Overweight and Obese Adults: An AI4Food Machine Learning Study
- Complexity analysis of physiological signals in the wearable era: New insights into chronic diseases and aging
- Personalized stress detection from wearable physiological signals via Digital Twin-inspired subject embeddings and recurrent neural networks
- Proteomic signatures of BMI generalize across ancestries and reveal cardiometabolic heterogeneity beyond measured BMI
- Cardiometabolic health-frailty phenotypes and bidirectional transitions between atrial fibrillation and heart failure: multistate trajectories and proteomic signatures.
BibTeX
@misc{kohn2026foundation,
title = {{A foundation model of wearable pulse oximetry reveals physiological signatures of health and cardiometabolic risk}},
author = {S. Kohn and G. Lutsker and A. Diament and S. Shilo and A. Gabet and G. Sasson and G. Wolf and A. Wolf and A. Godneva and A. Weinberger and H. Rossman and E. Segal},
howpublished = {medRxiv},
year = {2026},
month = jul,
doi = {10.64898/2026.07.01.26356992},
url = {https://www.medrxiv.org/content/10.64898/2026.07.01.26356992v2},
note = {CC BY},
}
RIS
TY - UNPB
TI - A foundation model of wearable pulse oximetry reveals physiological signatures of health and cardiometabolic risk
AU - S. Kohn
AU - G. Lutsker
AU - A. Diament
AU - S. Shilo
AU - A. Gabet
AU - G. Sasson
AU - G. Wolf
AU - A. Wolf
AU - A. Godneva
AU - A. Weinberger
AU - H. Rossman
AU - E. Segal
PY - 2026
DA - 2026/07/01
PB - medRxiv
DO - 10.64898/2026.07.01.26356992
UR - https://www.medrxiv.org/content/10.64898/2026.07.01.26356992v2
AB - While Photoplethysmography (PPG) is established as a noninvasive optical tool for monitoring heart rate and oxygen saturation, its high-resolution pulsatile blood volume waveforms contain rich physiological data that extend far beyond conventional vital signs. We introduce PulseOx-FM, a foundation model, trained using self-supervised learning on 6,995,558 segments of pulse oximetry signals collected during 42,282 overnight sleep monitoring recordings of 10,704 adult participants in the Human Phenotype Project (HPP). Using chronological age as a global health benchmark, PulseOx-FM significantly outperformed existing open-source and proprietary feature extraction methods while demonstrating robust generalization in an external out-of-distribution cohort. PulseOx-FM representations predicted 71 phenotypic targets spanning cardiometabolic, and neuropsychiatric domains beyond age, sex and body mass index. Applied in the same way to 3,721 intensive-care patients, two-minute windows of routine bedside pulse oximetry from the first day of admission improved the prediction of survival beyond age, sex and the heart rate and oxygen saturation the same device reports. Nightly embeddings further tracked next-day glycemic, dietary and activity-based state within individuals, dissociating this signal from sleep architecture alone. This next-day glycemic signal was predominantly a direct physiological effect, not explained by next-day dietary intake. These findings suggest that PulseOx-FM provides a generalizable framework for encoding physiological patterns from sleep, offering a non-invasive tool for global health risk stratification and precision medicine.
ER -