Today in @NatureMedicine we report that AI can predict 130 diseases...
James Zou@james_y_zou
43 views
Jan 07, 2026
~1 min read
Advertisement
1
Today in @NatureMedicine we report that AI can predict 130 diseases from 1 night of sleep🛌
We trained a foundation model (#SleepFM) on 585K hours of sleep recordings from 65K people—brain, heart, muscle & breathing signals combined.
AI learns the language of sleep🧵
We trained a foundation model (#SleepFM) on 585K hours of sleep recordings from 65K people—brain, heart, muscle & breathing signals combined.
AI learns the language of sleep🧵
2
Participants are linked to their EHR. SleepFM predicts risks for diverse diseases--including dementia, heart failure, kidney disease, and stroke--years before clinical diagnosis.
It substantially outperforms using demographic features, which are strong predictors. 2/n
It substantially outperforms using demographic features, which are strong predictors. 2/n
3
We created a new architecture to integrate multimodal sleep time-series data.
CNNs learn local features, transformers aggregate information across time + channels, and leave-one-modality-out contrastive learning trains robust representations.
This design generalizes across sites and diverse populations. 3/n
CNNs learn local features, transformers aggregate information across time + channels, and leave-one-modality-out contrastive learning trains robust representations.
This design generalizes across sites and diverse populations. 3/n
4
We spend 1/3 of our lives sleeping but it has been underexplored with AI.
Most work focuses on narrow tasks like sleep staging and apnea detection.
By learning a holistic representation of sleep, SleepFM opens new doors for studying the science and medicine of sleep. 4/n
Most work focuses on narrow tasks like sleep staging and apnea detection.
By learning a holistic representation of sleep, SleepFM opens new doors for studying the science and medicine of sleep. 4/n
5
Paper: nature.com/articles/s4159…
Open source code: github.com/zou-group/slee…
Truly wonderful collaboration with Emmanuel Mignot's lab, led by @connect_thapa and Magnus Kjaer! Thanks to all the awesome collaborators👏 n/n
Open source code: github.com/zou-group/slee…
Truly wonderful collaboration with Emmanuel Mignot's lab, led by @connect_thapa and Magnus Kjaer! Thanks to all the awesome collaborators👏 n/n
6
Our paper also discusses limitations of the study. In particular, participants in sleep studies differ from the general population and we don't make causal claims in the analyses. PSG recordings capture more signal than standard consumer wearables, which is an important future direction. There's much that we don't understand about sleep!