Perceptual and Motor Learning in Human Stick Balancing

16 We investigated the acquisition of skill in balancing a stick (52 cm, 34 g) on the fingertip in 17 nine participants using three-dimensional motion analysis. After 3.5 hours practice over 6 18 weeks, the participants could more consistently balance the stick for longer durations with 19 greatly reduced magnitude and speed of stick and finger movements. Irrespective of level of 20 skill, the balanced stick behaved like a normal non-inverted pendulum oscillating under 21 greater-than-gravity torque with simple harmonic motion about a virtual pivot located at the 22 radius of gyration above the center of mass. The control input parameter was the magnitude 23 ratio between the torque applied on the stick by the participant and the torque due to gravity. 24 The participants utilized only a narrow range of this parameter, which did not change with 25 practice, to rotate the stick like a linear mass-spring system. With increased skill, the stick 26 therefore maintained the same period of oscillation but showed marked reductions in 27 magnitude of both oscillation and horizontal translation. Better balancing was associated with 28 1) more accurate visual localization of the stick and proprioceptive localization of the finger; 29 and 2) reduced cross-coupling errors between finger and stick movements in orthogonal 30 directions, i.e., finger movements in the antero-posterior plane became less coupled with stick 31 tip movements in the medio-lateral plane and vice versa. Development of this fine motor skill 32 therefore depended on perceptual and motor learning to provide improved estimation of 33 sensorimotor state and precision of motor commands to an unchanging internal model of the 34 rotational dynamics. 35