Related Experiment Video
Updated: May 12, 2025
12:18
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
7.4K
A sleep-based risk model for predicting dementia: Development and validation in a Korean cohort
Hyukjun Lee1, Ji Won Han1,2, Seung Wan Suh3
1Department of Neuropsychiatry, Seoul National University Bundang Hospital, Seongnam, South Korea.
Journal of Alzheimer'S Disease : JAD
|May 8, 2025
Summary
A new Dementia Risk Score (DRS) effectively predicts dementia, including Alzheimer's disease, by incorporating sleep symptoms. This tool aids in early detection and community-based screening for cognitive decline.
Area of Science:
- Gerontology
- Neurology
- Public Health
Background:
- Dementia poses a significant public health challenge.
- Existing prediction models often neglect sleep-related symptoms, despite their established link to cognitive decline.
Purpose of the Study:
- To develop and validate a four-year Dementia Risk Score (DRS).
- To incorporate self-reported sleep symptoms with demographic and clinical factors for predicting all-cause dementia and Alzheimer's disease.
Main Methods:
- Analysis of data from 3082 Korean adults aged 60-79 years.
- LASSO regression for predictor selection, followed by multivariate logistic regression.
- Construction of a point-based DRS and internal validation using bootstrapping and a separate dataset.
Main Results:
- The DRS demonstrated robust predictive performance with AUC values of 0.824 (training) and 0.826 (validation).
- Significant predictors included sleep disturbance, sleep medication use, daytime dysfunction, leg discomfort, and urge to move legs.
Conclusions:
- The DRS offers a practical and scalable method for dementia risk prediction.
- It supports community-based screening and early intervention strategies for cognitive decline.
- Further external validation is recommended to confirm broader applicability.
