Abstract
Observers are sensitive to many sources of motion information. Studies have differentiated between the processing of first-order (luminance-defined) motion and motion defined by second-order cues such as contrast, texture, or disparity. Additionally, observers can use attention to track spatial displacement of features, reconstructing motion direction from sequential shifts in position. These disparate sources of motion information have different spatial and temporal characteristics, and consequently are not equally reliable across stimulus conditions. Typically, first-order motion information dominates perception at short durations. However, we previously demonstrated that increasing the size of high-contrast moving stimuli preferentially suppresses first-order motion information, while sparing second-order motion information (Glasser & Tadin, 2011, JOV). Here, we hypothesized that observers would make systematic errors in the discrimination of large, brief moving stimuli that may reflect changes in the strength and validity of different motion cues. Using a custom 360Hz display system, we found that, for brief stimuli (Gaussian temporal envelope, σ <15 ms), observers consistently perceive the motion direction of large, high-contrast gratings opposite to their physical movement (i.e., leftward motion is reliably perceived as rightward, and vice versa). If either stimulus size or contrast is decreased, observers’ discriminations again became veridical. These manipulations (increased size and contrast) are associated with stronger spatial suppression, and therefore with weakened first-order motion processing. Finally, we also show that motion reversal does not occur with non-periodic narrow-band random noise stimuli. These results suggest several, not necessarily exclusive explanations. The perceived motion reversal might arise from veridical motion information being suppressed by non-informative, or weakened, cues (cf. Baker, Meese & Georgeson, in press, i-Perception). Alternatively, in the absence of reliable first-order information, perception may rely on second-order and/or tracking mechanisms that may be more susceptible to illusory reversals or aliasing. We are currently conducting experiments to differentiate between these possible explanations.
Meeting abstract presented at VSS 2013