Abstract
Interocular suppression can be quantified using random-dot kinematograms (RDK) in dichoptic views, with coherently moving signal dots in one eye and randomly moving noise dots in the other. People with normal vision combine visual information into a fused image, so that signal detection is compromised by noise in either eye (Edwards, Badcock, and Nishida, Vis Res 1996; Hess, Mansouri, and Thompson, OVS 2010). However, the details of this combination are unclear. We used dichoptic stimuli in which signal and noise dots had different luminance contrasts in order to measure contrast-ratio thresholds in a 2AFC motion discrimination task. Observers had normal vision. They viewed stimuli through a stereoscope and reported the motion direction of the signal dots (up or down). A 3-down 1-up staircase controlled luminance contrasts. Signal and noise dots were independently divided between the eyes (0%, 25%, 50%, 75%, 100% to the left eye), which gave 25 dot-distribution conditions. There were three levels of global percent coherence (8%, 16%, 24%). We found that logCR threshold decreased with increasing percent coherence, and there was a small practice effect. The 25 conditions were grouped in several ways to measure characteristics of binocular combination. All observers slightly preferred binocular signal. Some observers had lower thresholds with signal and noise in different eyes, monocular noise, and monocular stimuli generally: they were able to exclude monocularly presented noise. Others preferred having signal and noise together in the same eye(s), binocular noise, and binocular stimuli generally: noise in one eye caused suppression of the other. All observers appear to have looked for signal within a binocularly combined representation, but they differed in their strategies for excluding noise. The data also revealed two regimes that observers used to do the task: reliance on global percent coherence or segregation of signal from noise by contrast grouping.
Meeting abstract presented at VSS 2013