Abstract
Crowding is characterized by an impaired ability to identify stimuli in the context clutter. Critical spacing, defined as the minimal distance by which stimuli must be separated to achieve criterion-level performance, increases with increasing eccentricity. Fixed-property models hold that critical spacing is determined by fixed properties of the visual system, such as fixed integration fields or fixed attentional resolution at given eccentricities. Fixed-property models, however, seem inconsistent with observations that when a target stimulus is distinguished from flanking stimuli by, for example, contrast polarity, crowding can be substantially reduced. These results suggest that representing target stimuli as objects that are distinct from surrounding stimuli may protect them from crowding because integration/selection processes are mediated through object representations. We tested this hypothesis by asking whether the object history of stimuli can protect a target from crowding even if there are no featural distinctions (e.g., contrast polarity) in the crowding display itself. Displays began with movies of identical gray squares moving along random paths. In one condition the squares moved coherently along a single path, thereby encouraging a representation of a single coherent, if complex, object. In the other condition, the squares moved along a separate paths, thereby encouraging representations of distinct objects. In both conditions, the motion ended with one square in the middle surrounded by four other squares and gaps appeared in one side of each square. Observers reported the gap position for the center square. Thus the displays in which the target was identified were identical across conditions; they differed only in their history and whether the center square had been perceptually encouraged to be represented as distinct from the surrounding squares or not. Less crowding was observed in the separate-motion condition than in the coherent-motion condition. These results rule out basic fixed-property models of crowding.
Meeting abstract presented at VSS 2013