JMIR Publications

ABSTRACT

Background:

Finger tapping tasks, which assess fine motor control, have emerged as potential digital biomarkers of cognitive function. With the widespread availability of smartphones, these tasks can be readily administered at home and other non-clinical settings. However, the relationship between smartphone-derived finger tapping metrics and cognitive performance remains underexplored.

Objective:

This study aimed to examine the association between smartphone digital finger tapping features and cognitive performance in an aging population.

Methods:

Participants were enrolled in the study as part of the electronic Framingham Heart Study (eFHS). They were instructed to perform a two-finger tapping task every two months over a one-year period. Each session lasted 20 seconds: participants alternated tapping with two fingers of the left hand for 10 seconds, followed by the right hand for 10 seconds. Cognitive performance was assessed using a standardized neuropsychological battery, evaluating global cognition and four domains: memory, executive function, language, and visuospatial function. Eighteen tapping features were extracted to capture aspects of motor performance, including the mean, median, standard deviation (SD), skewness, and slope of the inter-tap interval (ITI). Associations between tapping-derived features and cognitive outcomes were assessed using linear mixed-effects models, adjusting for age, sex, handedness, education, cohort, and the time interval between cognitive and tapping assessments.

Results:

A total of 302 participants (mean age = 74.7 years; 56.0% women, 13.2% non-White) with cognitive performance data completed the digital finger tapping tasks. Thirteen tapping features were significantly associated with global cognitive function (all P < .001). Each one SD increase in the number of taps was associated with a 0.14-unit higher global cognitive function score (β = 0.14, 95% CI: 0.07, 0.21). Furthermore, finger tapping features were significantly associated with multiple cognitive domains, with the greatest number of associations observed for executive function (13 significant features), including a strong association with mean ITI (β = –0.27; 95% CI: –0.33, –0.20). Stratified analyses by hand showed consistent effect directions across both hands. The aggregated composite scores derived from finger tapping features demonstrated significant associations with their respective cognitive test scores with correlation coefficient ranging from 0.30 for memory to 0.55 for executive function.

Conclusions:

Digital finger tapping features, particularly those reflecting ITI variability, are significantly associated with cognitive performance. These findings suggest that tapping tasks may serve as non-invasive, scalable tools for detecting cognitive decline, especially in domains of global cognition and executive function. Further research is warranted to validate their utility for early detection and longitudinal monitoring of cognitive health in aging populations.