abstract
- Contrary to a strict specificity of learning position, Robertson, Collins, Elliott, and Starkes (1994) have reported that the balance beam performance of expert gymnasts is less affected by the withdrawal of vision than is the performance of novice gymnasts. In this study, we employed a training paradigm in order to exercise complete control over the sensory conditions under which a dynamic balance beam task was acquired. Novice participants were trained either with or without vision to walk across a balance beam as quickly as possible and later tested in the other vision condition. Although participants improved more in the condition in which they trained, practice in one sensory condition did not negatively affect performance in a different sensory circumstance. The finding that vision was still extremely important after 5 days of practice is problematic for models of motor learning that propose a progression with learning from closed-loop to open-loop control.