Public Health

Area of Science:

• Neuroscience
• Cognitive Science
• Biostatistics

Background:

• The digital clock drawing test (dCDT) uses a digital pen to record pen movements for cognitive screening.
• Traditional dCDT analysis relies on summary features, which can be subjective and prone to missing data issues.
• Novel functional features, mathematically defined, are introduced to overcome these limitations.

Purpose of the Study:

• To introduce and evaluate novel functional features derived from the dCDT.
• To compare the predictive performance of these functional features against traditional summary features for cognitive impairment detection.
• To assess the robustness and objectivity of functional features in cognitive assessment.

Main Methods:

• Utilized dCDT data from 3,415 participants in the Framingham Heart Study.
• Trained random forest models with five-fold cross-validation to differentiate between cognitively impaired and intact individuals.
• Incorporated demographics, time-based, summary, and novel functional features (G-function, pressure density, radius function) into predictive models.

Main Results:

• Functional features measuring spatial proximity and circularity demonstrated high predictive accuracy (AUC: 0.90-0.91) comparable to traditional summary features.
• These functional features were robust when combined with demographics and time-based metrics.
• Similar predictive power was observed when analyzing command and copy tasks separately.

Conclusions:

• Functional dCDT features offer advantages like missing data robustness and reduced subjectivity.
• These features provide a more objective way to analyze complex dCDT data compared to summary statistics.
• Further research is needed to fully understand the cognitive insights provided by these functional features.