Predicting Cognitive Decline in Motoric Cognitive Risk Syndrome Using Machine Learning Approaches

Area of Science:

• Neuroscience
• Gerontology
• Biostatistics

Background:

• Motoric Cognitive Risk Syndrome (MCR) signifies a preclinical state for cognitive decline.
• Early prediction of cognitive deterioration in MCR is crucial due to variable progression.
• Identifying individuals with MCR who will progress is essential for timely intervention.

Purpose of the Study:

• To develop and validate a machine learning model for predicting short-term cognitive decline in MCR.
• To identify key predictors of cognitive decline in individuals with MCR.
• To assess the feasibility of integrating multimodal data for predictive modeling.

Main Methods:

• Utilized Support Vector Machine (SVM) classifiers to predict cognitive decline in 80 MCR participants (≥60 years).
• Baseline assessments included plasma biomarkers (β-amyloid, tau), dual-task gait, and neuropsychological tests.
• Models were trained and validated using feature importance analysis and cross-validation.

Main Results:

• Plasma β-amyloid and tau, dual-task gait, and memory scores were significant predictors of cognitive decline.
• The optimal linear SVM model achieved 88.2% accuracy and 83.7% AUC on the test set.
• Cross-validation demonstrated high performance with 95.3% average accuracy and 99.6% AUC.

Conclusions:

• Machine learning effectively integrates biomarker, motor, and cognitive data to predict cognitive decline in MCR.
• The developed model shows potential for clinical application in identifying at-risk individuals.
• External validation is recommended to confirm the generalizability and robustness of the predictive model.