Abstract
The increase of virtual reality tool use in rehab protocols makes the study of their shortcomings, such as system delays, crucial. For instance, introducing visual feedback delays has been shown to increase overshooting and deteriorate adaptation to visuomotor rotations in reaching and tracking tasks. Nonetheless, upon prolonged exposure, delay adaptation is possible, as reflected by improvements in accuracy and restored rotation adaptation. Here we tested whether simultaneous adaptation to two different delays is possible, when each is associated with a different target, as already shown possible for visuomotor rotation. We utilized a reaching task in which two opposite targets (90°, 270°) were each presented with a different cursor lag (0ms, 225ms). After a 240 trial habituation block, two types of probe trials involving a 30° cursor rotation presented either with a 0ms or 225ms lag, were introduced to measure post rotation biases (PRB). We reasoned that dual adaptation should entail both a gradual reduction in overshoots for the delay target and an interaction (cursor delayXtarget delay) when comparing PRBs. Surprisingly, participants persisted in overshooting the delay target in the habituation block (p=0.6). Moreover, only cursor delay affected adaptation, with smaller PRBs when cursor lagged for both targets (p=0.006) and no interaction (p=0.8). These data demonstrate that dual adaptation to visual feedback delays appears impossible, suggesting that the brain can only maintain a single relationship with feedback timing in a given task context. More generally, the challenge set by dual adaptation seems greater for temporal than spatial perturbations.
