Related Experiment Video
Updated: Jun 23, 2026
12:18
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Machine learning-based prediction model for cognitive frailty in elderly patients with ischaemic stroke: a
Xuan Chen1, Linjie Zhou2, Ying Zhang3
1School of Nursing, Air Force Medical University, Xi'an, Shaanxi, China.
Frontiers in Neurology
|June 22, 2026
Summary
Machine learning accurately predicts cognitive frailty (CF) in older stroke patients. The random forest model identifies key risk factors like stroke severity and age, aiding early intervention for better outcomes.
Area of Science:
- Neurology
- Geriatrics
- Artificial Intelligence
Background:
- Cognitive frailty (CF), a combination of cognitive impairment and physical frailty, is prevalent in older adults post-ischaemic stroke (IS).
- CF is linked to poorer functional outcomes in this vulnerable population.
- Early identification of CF risk is crucial for timely intervention.
Purpose of the Study:
- To develop and internally validate a machine learning (ML)-based model for predicting 3-month CF risk in older IS patients.
- To identify key predictors of post-stroke cognitive frailty.
Main Methods:
- A prospective cohort study enrolled 402 older IS patients.
- 26 baseline variables were assessed, and feature selection was performed using LASSO regression.
- Ten supervised ML algorithms were evaluated, with the Random Forest (RF) model selected for interpretability using SHAP analysis.
Main Results:
- 149 patients (37.1%) developed CF at 3 months.
- The RF model achieved an AUC of 0.889, accuracy of 0.798, and sensitivity of 0.909.
- Key predictors included discharge NIH Stroke Scale score, age, white matter hyperintensity, depression, and social support.
Conclusions:
- An interpretable ML model was developed to estimate early CF risk in older IS patients using routine clinical data.
- Neurological, nutritional, and psychosocial factors appear to interact in contributing to post-stroke CF.
- External validation is recommended before clinical implementation of the developed model.