Abstract
Visuomotor adaptation has been traditionally explained as the gradual update in an internal model based on a visual discrepancy between an intended movement and the resultant outcome. Under this framework, it has been assumed that the input to the motor system depends uniquely on a veridical estimate of the metrical properties of the scene. However, this assumption contrasts with many studies showing that action reflects biased estimates of 3D information. Whereas these distortions can cause problems for action, rapid motor adaptation helps overcome them. Yet, it remains unclear if adaptation itself relies on an accurate analysis of depth. Here, we tested whether motor learning processes were sensitive to changes in 3D information despite the same visual error which was being presented. Subjects reached to a visual target (a 6 degrees wide circle) that was either alone (No-depth condition) or surrounded by a virtual environment (Depth condition). To induce visuomotor remapping we imposed a task-irrelevant-error clamp: Regardless of the actual hand's direction, the cursor always hit off the target by a fixed horizontal offset. Although this visual error was irrelevant for the task, subjects displayed robust adaptation: They moved in a direction opposite of the visual error induced by the clamp. Furthermore, this motor learning was stronger in the No-depth condition compared to the Depth condition. Remarkably, these results mirrored those of a perceptual task, where subjects adjusted distance and width of a probe to match the target's perceived location and size under the same two visual conditions. That is, the degree of adaptation was sensitive to perceived depth. These results demonstrate that visuomotor adaptation takes into account a more sophisticated error signal than previously thought, integrating also sources of information that are irrelevant to the task
Meeting abstract presented at VSS 2018