Abstract
Real-world size is an intrinsic property of objects: it is accessed automatically when we see an object and predicts a consistent medial-to-lateral organization in ventral temporal cortex (Konkle & Oliva, 2012). Is real-world size a purely abstract concept or a conceptual distinction reflected in perceptual differences? Since visual search performance is primarily determined by perceptual similarity, if big and small objects tend to be perceptually different from each other, it should be easier to find a target object among distractors with a different (vs. same) real-world size than the target. We examined visual search rates for big and small manmade objects (Experiment 1) and big and small animals (Experiment 2). The stimulus sets were matched for familiarity, typicality, aspect ratio, area, contour spikiness, object extent; gray-scale images were equalized across luminance, contrast, and power spectrum (spatial-frequency/orientation). On each trial, a target item was previewed, followed by a search display with 3 or 9 items. All items were presented at the same physical size on the screen, but the distractor items were either from the same or different real-world size category as the target. Participants pressed the space bar when they found the target, after which all items were replaced with X’s and participants clicked on the target’s location. In both experiments, participants were significantly more efficient at finding the target item among different real-world size distractors than among similar real-world size distractors, as indicated by faster search slopes for different-size trials (Objects: different-size slope: 51.88 ms/item, same-size slope: 61.03 ms/item, p<.01; Animals: different-size slope: 70.26 ms/item, same-size slope: 76.88 ms/item, p<.05). Because we controlled for differences in a wide range of basic feature and shape dimensions, these results suggest that real-world size is a conceptual distinction that correlates with mid/high-level perceptual features that guide visual search.
Meeting abstract presented at VSS 2013