Research
Deep learning-based gait analysis predicts multisystem health phenotypes
Nature ·
The Brief
- A gait foundation model was built from 3D skeletal motion data of 3,328 deeply phenotyped adults.
- Motion was recorded with a depth camera during five motor tasks.
- The study was published in Nature Medicine on 9 November 2026 (doi 10.1038/s41591-026-04725-8).
- According to the source excerpt, embeddings from the model significantly predict multisystem health phenotypes.
Why it matters
It suggests that routine, camera-based gait measurements could serve as a scalable window into multiple aspects of health, which is relevant to researchers in digital biomarkers, clinical measurement and medical AI.
Who’s affected
- clinical researchers and physicians interested in digital biomarkers
- medical AI and computer vision researchers
- gerontology and public health researchers
- #gait analysis
- #deep learning
- #digital biomarkers
- #foundation models
- #multisystem health
Nature Medicine, Published online: 09 November 2026; doi:10.1038/s41591-026-04725-8 We developed a gait foundation model from three-dimensional skeletal motion data of 3,328 deeply phenotyped adults, recorded using a depth camera during five motor tasks. Embeddings significantly…
Timeline
Study published online in Nature Medicine
nature.com
Deep learning-based gait analysis predicts multisystem health phenotypes