A Machine Learning Analysis of The Bead Maze Hand Function Test for Predicting Manual Dexterity in Children.

Comprehending how forces are applied to an object during manipulation can help provide important insights into the quality of behavior in daily tasks. We have developed the Bead Maze Hand Function test to objectively measure the quality of hand function in children. This test aims to measure how well an individual performs the activity by integrating measures of time and force control. The main objective of this study was to examine associations between a common clinical measure, the Box and Block Score (BBS), and variables on the Bead Maze Hand Function test that were either time-based or force-based. The sample was composed of neurotypical developing adolescent participants (N=23). We found that the time (duration) on the double curve wire (most complex) was the best predictor of BBT. Furthermore, we found that force-based variables were weak predictors of the clinical, time-based BBS. These findings support the integration of time and force-based metrics to holistically quantify the quality of motor behavior. Linking these metrics into a unified score may serve as a better way to analyze adolescent motor behavior and predict future motor or neurodegenerative conditions.