Abstract
We investigated how interactions between monocular motion parallax and binocular cues to depth vary in human motion areas for wide-field visual motion stimuli (110 × 100°). We used fMRI with an extensive 2 × 3 × 2 factorial blocked design in which we combined two types of self-motion (translational motion and translational + rotational motion), with three categories of motion inflicted by the degree of noise (self-motion, distorted self-motion, and multiple object-motion), and two different view modes of the flow patterns (stereo and synoptic viewing). Interactions between disparity and motion category revealed distinct contributions to self- and object-motion processing in 3D. For cortical areas V6 and CSv, but not the anterior part of MT+ with bilateral visual responsiveness (MT+/b), we found a disparity-dependent effect of rotational flow and noise: When self-motion perception was degraded by adding rotational flow and moderate levels of noise, the BOLD responses were reduced compared with translational self-motion alone, but this reduction was cancelled by adding stereo information which also rescued the subject's self-motion percept. At high noise levels, when the self-motion percept gave way to a swarm of moving objects, the BOLD signal strongly increased compared to self-motion in areas MT+/b and V6, but only for stereo in the latter. BOLD response did not increase for either view mode in CSv. These different response patterns indicate different contributions of areas V6, MT+/b, and CSv to the processing of self-motion perception and the processing of multiple independent motions.
Highlights
Depth can be perceived with a single eye when an object approaches
Is there a general preference for one of the two self-motion types (RT and T motion, sm data only), irrespective of view mode? We found a strong preference for Rotation+Translation flow (RT) motion in CSv, p2v, and ventral intraparietal area (VIP) [RM ANOVA, F(1, 11) = 24.51; 25.01; 6.91, p < 0.05, see Figure 7B]
Our results show that the response pattern in V6 does not exhibit a simple self-motion preference, but that its motion type preference is dependent on the presence of disparity signals
Summary
Depth can be perceived with a single eye when an object (or a whole scene) approaches. The visual flow of the approaching scene gives rise to relative motion between the images of nearer and farther elements of the scene: motion parallax. Depth cues from motion parallax and stereo interact most convincingly in area V3b/KO (Ban et al, 2012). For natural movement through the world, depth analysis is often far more complicated. Independently moving elements violate the property that the motion-parallax cue to depth depends on a rigidly moving scene or object relative to the observer. Moving elements make relative motion ambiguous: it can be caused by object movement in the world, by depth or both
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
