One Night's Sleep May Predict 130 Diseases - Takeaways - MDSpire

One Night's Sleep May Predict 130 Diseases

Foundation model trained on more than half a million hours of polysomnography data outperforms demographic baselines in forecasting mortality, dementia, and cardiovascular outcomes

  • By

  • Kerri Miller

  • January 8, 2026

  • 6 min

Share

  • 1

    The SleepFM model predicts 130 diseases from a single night of sleep data with a C-Index of at least 0.75 for all conditions.

  • 2

    SleepFM was trained on over 585,000 hours of polysomnography data from approximately 65,000 patients across multiple cohorts.

  • 3

    The model demonstrated strong predictive performance for conditions like dementia (C-Index 0.85) and myocardial infarction (C-Index 0.81).

  • 4

    SleepFM outperformed baseline models, achieving significant accuracy improvements in predicting all-cause mortality and neurological conditions.

  • 5

    The model also performed well on standard sleep analysis tasks, achieving competitive scores in sleep staging and sleep apnea classification.

Original Source(s)

One Night's Sleep May Predict 130 Diseases

  • by Kerri Miller

  • January 8, 2026

Related Content

Upfront Cost Guidance Limited in US Guidelines

Although most large US-based medical societies in a cross-sectional study explicitly suggested considering economic or resource use–related evidence during guideline development, only 3.8% of individual recommendations included that information in the recommendation statement itself.

Serostatus May Modify RA Activity–CV Link

Combined rheumatoid factor and anticitrullinated protein antibody status appeared to modify the association between baseline rheumatoid arthritis disease activity and subsequent cardiovascular events in an international observational cohort.

Prolonged Sleep Restriction May Increase Body Weight

In a pooled analysis of two randomized crossover trials, reducing nightly sleep by about 1.5 hours for 6 weeks was associated with modest increases in body weight and waist circumference without measurable changes in body composition.