Abstract
Extended reality (XR) systems show promise for upper-limb rehabilitation, yet the influence of task design parameters on tracking performance remains poorly characterized. This pilot study (
n
= 10 healthy adults) investigated how the size and number of virtual targets affect tracking performance in an XR fingertip-tracking task. Specifically, we examined sphere diameter (2, 5, and 8 cm) and target count (10, 15, and 20) across four 3D path geometries (Straight, S-curve, Circle, and Zigzag). Mean tracking error and task completion time were analyzed using linear mixed-effects models with participant ID as a random intercept and Kenward–Roger degree-of-freedom adjustments. Sphere diameter emerged as the most consistent predictor: smaller targets were associated with lower tracking error but longer completion times across all geometries. In contrast, target count showed largely non-significant effects. The Circle geometry exhibited a distinct performance pattern relative to the other shapes. These pilot findings inform task-parameter selection for XR-based upper-limb rehabilitation and motivate a larger, pre-registered study.