Poster Feature development of digital measures for the finger tapping task

Feature Development of Digital Measures for the Finger Tapping Task

Koneksa’s finger tapping digital biomarker poster looks at smartphone-based hand dexterity data to identify which derived features actually track Parkinson’s disease symptoms.

The Finger Tapping task asks patients to alternately tap two on-screen buttons, recording the coordinates and timestamp of every tap. From that raw data, Koneksa derived six candidate features and tested each one against two groups from the mPower study: self-stated Parkinson’s patients in an off-state, and self-stated healthy participants.

Each feature’s distribution was compared across cohorts using the Kolmogorov-Smirnov test, with a p-value threshold of less than 0.05 used to identify meaningful differences.

Three features: tapping speed, sequence correctness, and speed regularity — distinguished the groups with strong significance, down to p-values of 6×10⁻⁹. Two others, left-right speed asymmetry and change in tapping speed, showed no significant difference. A sixth feature, positional accuracy, was also significant, but in an unexpected direction: off-state subjects were more accurate than healthy subjects.

This finger tapping digital biomarker poster reinforces a broader principle in digital measure development: not every computable metric is clinically useful, and validating features against retrospective data is an essential step before building a digital endpoint.

View the poster to see the full results, or contact us to discuss digital assessment design for your Parkinson’s program.