JMIR Publications

ABSTRACT

Background:

The Clock Drawing Test has been widely used in clinic for the cognitive assessment. Recently the digital Clock Drawing Test (dCDT) was introduced that is able to capture the entire sequence of clock drawing behaviors. While a variety of domain specific features can be derived from dCDT, how they actually correlate with these different cognitive skills has not been evaluated in a large community-based population.

Objective:

This paper aims to investigate the association between dCDT features and the cognitive performance across multiple domains in the Framingham Heart Study (FHS), a large community-based cohort with longitudinal cognitive surveillance. Another aim is to investigate the diagnostic performance of dCDT features in classifying a subset of clinically diagnosed participants with mild cognitive impairment (MCI) from those with normal cognition (NC).

Methods:

The current study included dementia-free participants from the FHS who were administered both the dCDT and a standard protocol of neuropsychological (NP) tests that measure a range of cognitive functions. A total of 105 features were derived from dCDT, and correlated with 18 NP tests by linear regression models. We further built composite scores from dCDT features that were significantly associated with NP tests. The classification models were further built to distinguish a subset of participants that had been clinically diagnosed with MCI from those with NC.

Results:

Our study sample included 2,062 participants (mean age 62±13 years, 51.6% women), including 36 MCI cases. Each NP test was associated with an average of 50 dCDT features. All dCDT composite scores were significantly associated with MCI. The scores also showed superior performance in the classification of MCI than standard NP scores (0.81-0.85 vs 0.76-0.85).

Conclusions:

The dCDT performance is psychometrically valid to detect MCI in a large community-based cohort.