Abstract
The perception of coherent objects depends on grouping local image features into global spatial forms. Numerous cues support visual grouping, including similarity, spatial proximity, and common temporal structure, but knowledge of how the visual system combines these various sources of information remains limited. The current study investigates the extent to which grouping can occur without awareness of the cues that gave rise to the perceived organization. Observers viewed arrays of Gabor patches in which the spatial frequency of the elements changed stochastically over a one second trial. Within the array, a randomly positioned rectangle, oriented horizontally or vertically, grouped separately from the background elements. On some trials, this rectangle was defined by a temporal structure cue alone: all elements within the array were randomly oriented, but the figure elements followed one point process, while the background elements followed a different point process. On other trials, the rectangle was defined by a similarity cue alone: all elements within the array changed simultaneously, but the orientations of the figure elements were similar to one another, and differed from the orientations of the background elements. In a third condition, both temporal structure and similarity defined the figure. In several experiments, observers reported the orientation of the perceived figure, and judged which cue or cues triggered the organization. The results suggest that although both similarity and temporal structure strongly supported grouping, observers could not reliably identify the grouping cues; judgments of the temporal structure cue as present correlated more strongly with perceived figure strength than with the actual presence of the cue. These findings will be discussed in the context of ongoing discussions concerning potential mechanisms for stimulus binding.
Meeting abstract presented at VSS 2014