Abstract
Normally we integrate ordinal (occlusion) and metric (e.g., binocular disparity) depth information to obtain a unified percept of 3D layout. Further, quantitative depth must be available to the proprioceptive and motor systems to support interaction with nearby objects. Here we take a step towards understanding how occlusion and binocular disparity combine in the control of visually-guided reaching. We developed a novel conflict paradigm set in a real-world environment in which participants placed a virtual ring around a post positioned at one of several distances (34, 41.5, and 49 cm). The ring was fixed to the index fingertip (with lateral and vertical offsets to avoid finger collisions with the post). If the ring collided with the post, the ring changed colour and the trial restarted. We assessed performance with monocular and binocular viewing using both virtual and physical posts (N=10). The consistency of the occlusion was manipulated such that when the post was physical it never occluded the ring, even when correctly positioned around the post. This resulted in conflicting disparity and occlusion information between the post and further portion of the ring. Conversely, in virtual post conditions, occlusion of the ring by the post was consistent. We found that ring placement was less precise when occlusion and disparity information were inconsistent. Participants also required more attempts to complete the task under the conflict compared to consistent conditions. While this pattern of results was similar for binocular and monocular viewing, observers performed worse and required more attempts when doing the task with one eye. Our results underscore the importance of binocular depth information in performing visuo-motor tasks. However, even when such precise quantitative depth information is available, ordinal depth cues can significantly impact both perception and action, despite these latter cues only providing binary signals to the success of visually-guided action.