IMU-based linear and nonlinear metrics for the characterization of fine motor development in early childhood

Background: Fine motor competence assessment plays a pivotal role in neuropsychological examinations for the identification of developmental deficits. The Placing Bricks (PB) test evaluates fine motor competence relying on the measurement of time to completion. Recently, the PB test was instrumented using wearable inertial sensors to complement its standard assessment with reliable and objective measures of performance (i.e., cycle and placing durations). This work aims at extending the metrics that allow a quantitative fine motor assessment and at investigating fine motor trajectory from 3 to 7 years of age. Research question: How can IMU-based linear and nonlinear metrics provide insight into fine motor development during early childhood? Methods: 126 typically developing children (3-7 years) performed the PB test with inertial sensors on both wrists. Temporal parameters (Cycle and Placing durations), their variability (interquartile range, short-term and longterm variability, via Poincare plots) were calculated on angular velocities, and nonlinear metrics (multiscale entropy and recurrence quantification analysis) on acceleration data. The effects of age, handedness, and sex were analysed: parameters showing age effects were visualized using a polar plot with age reference bands. Results: Results revealed differences related to age and hand dominance: cycle and placing durations as well as variability decreased, while multiscale entropy increased and recurrence decreased with age and when using the dominant hand. Both age and the use of the dominant hand are associated with shorter, less variable cycles, with more complex movement patterns, reflecting more mature fine motor control. Significance: The patterns highlight specific characteristics of fine motor maturation, showing in which domain changes occur and towards which direction. The proposed IMU-based approach effectively described fine motor control development in the analysed population.