Abstract
The human visual system extracts disparity information from multiple sources, classified as first (luminance modulation), second (contrast modulation), and third (pattern/features) order. When these sources coexist, first order disparities are believed to play a dominant role. Here we present evidence to the contrary. We measured short-latency disparity vergence responses (DVRs) elicited in humans by the sudden presentation of binocular compound gratings, obtained by summing two or more sinusoidal gratings. Individual components were either all horizontal (with vertical disparity) or all vertical (with horizontal disparity). Appropriate combinations of spatial frequency (SF) and disparity yielded stimuli in which first, second, and third order signals coexisted, but had different SFs and disparities (of either sign). We found that with these stimuli the strength of the response to each signal is mostly determined by its own SF and contrast, just as when they are presented in isolation. Thus, the relative response strength to the three signals can be altered simply by scaling the stimulus. For example, in a 3F+5F stimulus first order dominates when the frequency of the fundamental F is low, and higher order dominates when it is high. At intermediate scales no one disparity signal dominates, and responses of similar magnitude to each type of disparity can be observed. However, because of the different latency of these responses (first order being the fastest, second order appearing 20ms later, and third order even later), they peak at different times. An exception to this rule is that at very low contrasts only first order disparities elicit DVRs. In conclusion, we found that in stimuli that contain multiple sources of disparity, DVRs simply reflect the relative strength of the various signals, determined by their individual SF and contrast. We found no evidence that first order signals play a privileged role.
Meeting abstract presented at VSS 2014