Sleep patterns in dementia differed from typical aging
After risk scores were grouped, the model identified 75.7% of dementia cases and correctly classified 74.7% of controls using data not used for training.
Our read
Biomarkers- Studied:
- Cohort study of 1,672 people from the general population and a dementia cohort, against controls
- What changed:
- The model identified 75.7% of dementia cases and correctly classified 74.7% of controls using unseen data
- The catch:
- The comparison with clinical judgment was a pilot in only 50 high-risk people
Editor of record: Boris Djordjevic · AgingResearch.News
People from the general population and a dementia cohort, totaling 1,672 participants, contributed under-the-mattress sleep sensor data to an observational study. Researchers developed a computer model to estimate age from repeated sleep measurements, then used deviations from typical age-related patterns to assess dementia risk. They grouped risk scores into high, medium and low categories.
Sleep patterns in people with dementia departed from typical aging patterns. These deviations were associated with irregular times for going to bed and getting up, and with less night-to-night variation in deep sleep. On data held back from training, the model’s age estimates were off by an average of 5.52 years.
Why it matters
For dementia risk screening, the authors suggest that remote sleep monitoring may help identify people who could benefit from further clinical evaluation.
Caveats
Separating dementia cases from controls is not the same as predicting who will develop dementia. The comparison with clinical judgment was a pilot involving 50 high-risk people.
The paper
UK Dementia Research Institute; Great Ormond Street Hospital
NPJ Digital Medicine, Jul 21, 2026, Open access
Comments
Be the first to comment.