Contribution of retinal and non-retinal sources of visual information in egocentric object localisation for action

In visuo-manual tasks, enhancing visual information by presenting a target-object on a textured background has been proved to improve distance coding. However, controversial interpretations have suggested that enriched retinal inputs either provides access to more veridical spatial information or enables extra-retinal signals (e.g. vergence) to be more accurately processed. To contrast these interpretations, the present study evaluated the spatial accuracy of upper limb movement towards visual targets presented monocularly or binocularly either in darkness or on a textured background. Action was performed whilst wearing neutral or prismatic spectacles deviating by 5.4° gaze direction but leaving spatial information on the retina virtually unchanged. We found that distance underestimation observed in darkness reduced in presence of a textured background whatever the vision condition. Surprisingly, the performance was not affected by prismatic displacement when a textured background was provided. These results indicate that distance processing for action relies mainly on retinal inputs in presence of a structured workspace, even when there is a mismatch between retinal and extra-retinal signals.