Abstract
Previous studies have shown that humans can discriminate angular velocities of rotations, and speeds and directions of translations in optic flows. These types of motions are two of the few components that one can decompose optic flows generated by small rigid patches of surfaces. A third important component is radial motion. We wanted to test whether its parameters, namely, the rate and focus of expansion can be measured by the visual system. To test the discriminability of the rate of expansion, we used a two-alternative forced-choice paradigm, controlling for local-velocity cues by using solid disks of different radii. In turn, to test for the effects of the mislocation of the focus of expansion, we manipulated its position relative to the fixation point in an expanding random-dot field. The results showed that rates of expansion could be finely discriminated regardless of the relative sizes of the stimuli. The Weber fraction for this discrimination was about xxx%. In contrast, if a large field of expanding dots has its focus of expansion displaced from the fixation point, then subjects overestimate the rate of expansion by as much as 20%. However, when positions around the fixation point are masked, this overestimation is reduced. Moreover, when the expansion field is small, the overestimation disappears. In conclusion, humans can discriminate rates of expansion finely, but have a rougher time locating the focus of expansion. It appears to be positioned near fixation, unless the expansion borders are small, serving as clue for positioning. When it is wrong, the visual system slightly miscalculates the rate of expansion. These results show that the measurements of the parameters of expansion and rotation are similar, suggesting almost identical mechanisms.
Supported by National Eye Institute Grants EY08921 and EY11170.