Missing depth cues in virtual reality limit performance and quality of three dimensional reaching movements

Goal-directed reaching for real-world objects by humans is enabled through visual depth cues. In virtual environments, the number and quality of available visual depth cues is limited, which may affect reaching performance and quality of reaching movements.

We assessed three-dimensional reaching movements in five experimental groups each with ten healthy volunteers. Three groups used a two-dimensional computer screen and two groups used a head-mounted display. The first screen group received the typically recreated visual depth cues, such as aerial and linear perspective, occlusion, shadows, and texture gradients. The second screen group received an abstract minimal rendering lacking those. The third screen group received the cues of the first screen group and absolute depth cues enabled by retinal image size of a known object, which realized with visual renderings of the handheld device and a ghost handheld at the target location. The two head-mounted display groups received the same virtually recreated visual depth cues as the second or the third screen group respectively. Additionally, they could rely on stereopsis and motion parallax due to head-movements.

All groups using the screen performed significantly worse than both groups using the head-mounted display in terms of completion time normalized by the straight-line distance to the target. Both groups using the head-mounted display achieved the optimal minimum in number of speed peaks and in hand path ratio, indicating that our subjects performed natural movements when using a head-mounted display. Virtually recreated visual depth cues had a minor impact on reaching performance. Only the screen group with rendered handhelds could outperform the other screen groups. Thus, if reaching performance in virtual environments is in the main scope of a study, we suggest applying a head-mounted display. Otherwise, when two-dimensional screens are used, achievable performance is likely limited by the reduced depth perception and not just by subjects' motor skills.