Abstract
Visual crowding is the impairment in peripheral object recognition due to clutter. Recent work has suggested crowding occurs not only between low-level features of stimuli, but between higher-level, holistic representations as well. Inverting the surrounding or flanking faces reduces the effect of crowding, suggesting a holistically-driven interference that contributes to the crowding observed with upright face flankers. However, it is not yet clear whether this flanker inversion effect is specific to faces or generalizes to other object categories. In a series of experiments, we asked participants to perform a discrimination task on either faces or cars, presented in the form of two-tone "Mooney" images lacking readily identifiable parts. Targets were presented at varying eccentricities, either alone or surrounded by upright or inverted flankers. Crowding was observed in the periphery for both types of stimuli in all experiments. In one experiment, we replicated the face flanker inversion effect when participants performed a gender discrimination task on face targets. In another experiment, we did not find an inversion effect for a left-right orientation discrimination task on car targets. Because the lack of an inversion effect for cars could be attributed to differences between task demands, a third experiment assessed crowding of both faces and cars using the same left-right orientation discrimination task. We observed the inversion effect for faces in the gender discrimination task, but found no inversion effect for either faces or cars during the orientation-discrimination tasks; therefore, our results suggest that crowding between holistic face representations may occur selectively for tasks that rely on holistic processing. Additional experiments involving tasks that engage holistic as opposed to strictly part-based mechanisms for processing of non-face objects are necessary in order to determine whether interference between higher-level representations is unique to faces.
Meeting abstract presented at VSS 2016