Dynamic Thumb Proprioception: Quantification with a Novel Robotic Task

The thumb plays a crucial role in hand function, yet its proprioceptive abilities remain poorly understood. Here we quantified dynamic thumb localization ability in unimpaired participants, using a novel task in which a robot moved the thumb in a circle and participants pressed a button when they felt their thumb aligning with a target point on a screen. After pressing the button, they received visual error feedback in the form of a ball jumping toward the target. To characterize thumb localization ability, we varied thumb speed and rotation diameter, assessed the effect of a propriovisual rotational perturbation, and compared index finger performance. Following task familiarization, thumb localization error was ~25° and did not change significantly with speed or circle diameter. Reversing thumb rotation increased error followed by rapid error adaptation across the next 20 trials, as would be expected if individuals formed an internal model based on a body-centered (movement-aligned) frame of reference rather than a world-centered spatial frame. Localization error was larger for the thumb than the index finger error for the same task (p = 0.02) and was correlated with a different, robotic assessment of finger proprioception (ρ = 0.61, p = 0.001). These findings indicate that dynamic thumb localization is somewhat inaccurate, although it can leverage visual feedback within a body-centered reference frame, a form of passive, cross-sensory adaptation. Further, in unimpaired adults, the dynamic proprioceptive abilities of the thumb and index finger are related, with thumb proprioception ability being less accurate than the finger.