This proposal investigates the potential benefits and limitations imposed by saccadic gaze shifts on surface slant perception, sensed with binocular disparity cues. Slant percepts are based on several retinal and extra-retinal cues. Retinal cues include horizontal and vertical disparity patterns and retinal eccentricity. Extra-retinal cues include information about version and vergence eye position. Their uncertainty or noise levels limit the accuracy with which these cues are represented. The amount of uncertainty associated with retinal cues varies with retinal image locus. Surface slant is sensed effortlessly when targets are imaged near the high-resolution region of the fovea. Slant is more difficult to estimate when targets are imaged in the periphery such as when comparing slant of widely separated targets and only one target can be imaged on the fovea at a time. Gaze shifts could facilitate slant resolution in several ways. They could improve resolution of retinal cues of widely spaced targets by imaging them sequentially on the fovea near the horopter. They could also provide extra-retinal signals for distance and direction of targets that are used to sense surface slant relative to the head. However, eye movements might also limit slant resolution if uncertainty about version and vergence eye movements during gaze shifts was greater than uncertainty of vertical disparity patterns and retinal image position of targets viewed peripherally with a stationary eye. Eye movements could also introduce temporal limitations for slant discrimination by shortening viewing duration, and introducing time delays between sequentially viewed stimuli. This proposal investigates how the visual sensory and motor systems interact using saccadic gaze shifts to reduce various sources of noise and improve the quality of retinal disparity cues for slant. This noise reduction should expand the spatial range of high-resolution stereo-slant perception.
Showing the most recent 10 out of 29 publications
