Predicting Global Cognitive Decline in the General Population Using the Disease State Index

Area of Science:

• Neuroscience
• Gerontology
• Biostatistics

Background:

• Early identification of individuals at risk for cognitive decline is essential for timely dementia detection and intervention.
• Predictive models can help select individuals who would benefit most from therapeutic or preventive strategies for dementia.

Purpose of the Study:

• To validate the Disease State Index (DSI), a supervised classification method, for predicting cognitive decline in the general population using multivariate data.
• To assess the predictive performance of DSI with various baseline features.

Main Methods:

• Included 2,542 non-demented participants from the population-based Rotterdam Study (mean age 60.9).
• Defined significant global cognitive decline as the top 5% of participants with the largest annual decline.
• Trained DSI using baseline features including MRI, cardiovascular risk factors, APOE-ε4, gait, education, and cognitive function; assessed performance using Area Under the Curve (AUC) via cross-validation.

Main Results:

• DSI achieved a mean AUC of 0.78 (0.77-0.79) using only age as a predictor.
• Incorporating all available features resulted in a slightly lower mean AUC of 0.77 (0.75-0.78).
• Models without age or with age-corrected features yielded a mean AUC of 0.70 (0.63-0.76), indicating age's strong predictive power.

Conclusions:

• The DSI model demonstrated the highest predictive performance for global cognitive decline when solely utilizing age as the input feature.
• While other features showed potential, they did not enhance prediction accuracy beyond age alone.
• Future research should explore novel features, such as longitudinal data, and alternative prediction methodologies to potentially improve predictive performance.