Abstract
Prediction turns motor commands into expected sensory consequences, whereas control turns desired consequences into motor commands. Flanagan and colleagues (2003) have shown that subjects can learn to predict before they can actually learn to control. This observation was interpreted as evidence that the update of prediction precedes control in motor learning. Here we investigated the transfer of learning between two visuomotor tasks both requiring adaptation to a 90° rotation. In the first task participants had to track with their eyes a self-moved target whose displacement was driven by random hand motion (see also Landelle et al., 2016). In the other task participants had also to move the hand but this time they were required to move a cursor so as to track an externally moving target (see also Ogawa & Imamizu, 2013). The first task was designed to test the ability of participants to predict novel visual consequences arising from their hand actions (eye tracking task). The second task was designed to monitor their ability to control a cursor along a desired trajectory (hand tracking task). Our preliminary results suggest an asymmetrical transfer of learning between the two tasks. Namely, although prior experience in the hand tracking task enhanced performance in the eye tracking task, prior experience in the eye tracking task did not improve performance in the hand tracking task. A possible scheme to account for these results is that visuomotor adaptation in our hand tracking task requires both the update of a forward and inverse model (Wolpert & Kawato, 1998), whereas adaptation in our eye tracking task relies solely on the update of a forward model. At a more general level these results emphasize that our ability to predict sensory consequences of hand movements can be improved without necessarily improving our ability to control hand movements.
Meeting abstract presented at VSS 2018