Emergent flexibility in motor learning.

We examined the effect of exploring redundant solutions during practice in enhancing the ability to flexibly use them to achieve a task goal. Three groups used different degrees of path redundancy to perform a virtual interception task in which they attempted to hit a stationary target by moving around a stationary obstacle. The low-variability group always practiced with the same position of the obstacle on all trials. The medium-variability and high-variability groups practiced with the obstacle in different positions within a range of 1 and 2 cm respectively. After eight blocks of practice, all participants were transferred to two tests: (a) a fixed obstacle test where the condition was the same as that practiced by the low-variability group, and (b) a variable obstacle test where the condition was the same as that practiced by the high-variability group. Results showed that the low-variability group had the most accurate performance both in the fixed obstacle and the variable obstacle test. The low-variability group showed the least path variability during the fixed obstacle test but was also able to adapt to the different positions of the obstacle during the variable obstacle test. It appears that flexibility in interceptive tasks is emergent from learning a particular task-relevant parameter related to the target location.