Abstract
Learning a motor skill requires production of qualitatively appropriate movement that can be fine-tuned (Newell, 1991) but children with disabilities, such as developmental coordination disorder (DCD), often cannot achieve this qualitative success independently (Clark et al., 2005; Smits-Engelsman et al., 2001). Passive modeling has been used to help learners accomplish the necessary approximation, but this approach frequently fails to work well (Kwakkel et al., 2008). Action theory suggests learners need to actively generate movements to be successful (e.g. Bingham, 1988) but it can be difficult to improve motor performance when reliable approximations of a target action are not available. Snapp-Childs et al. (2010) developed a method that enables children with DCD to overcome this ‘catch-22’ and improve manual performance to match typically developing children. In this study, we conducted a strong test of the role of active versus passive motor performance in the improvements found in Snapp-Childs et al. (2010). Fourteen adults completed 12 baseline trials where the task was to use a stylus (Phantom Omi) to push a bead quickly along a wire path in a 3D virtual environment; in these trials, magnetic attraction to the wire path systematically decreased thus increasing task difficulty. Half of the participants then completed 3 sessions of 36 active trials with varying levels of attraction, while the other half were given comparable passive experience manually tracking the trajectories. After training, both groups repeated the baseline and 4 novel trials (with longer, more complex, paths). Before training, there were no differences between groups in trial duration. The active group, however, performed significantly better than the passive tracking group after training (12.1 s vs. 15.4 s, p < 0.05). This effect was strongest in the novel trials (26.7 s vs. 40.8 s, p < 0.05) that required generalization. Active control during training aids generalized motor learning.