Broad category membership guides visual attention in young children

Bria Long, Mariko Moher, Talia Konkle, George A. Alvarez, Susan Carey

Introduction: When adults recognize objects, they automatically access knowledge about the broad category distinctions between animate vs. inanimate entities (animacy) and between big objects vs. small objects (object size; Konkle & Oliva, 2012). In addition, adult’s perceptual systems are sensitive to visual shape features that can distinguish inanimate from animate objects and big objects from small objects (Long, Konkle, Cohen, & Alvarez, submitted). How much experience is required for these broad category representations to emerge in the visual system? On one hand, broad category membership requires generalization across diverse basic-level categories (e.g., ants and zebras), arguably a computationally challenging task. Yet, as infants first group objects at broader levels (Mandler & Bauer, 1988), young children may already be sensitive to the visual features that distinguish animate from inanimate objects and big objects from small objects.

Methods: We used visual search to explore which category distinctions can guide visual attention in young children: if two categories tend to be perceptually different, it is easier to find a target from one category among distractors from another category versus the same category (Duncan & Humphreys, 1989). We conducted three visual search experiments with 3- and 4-year-old children using an iPad, testing the broad categorical distinctions of animacy, object size, and edibility. Stimuli spanned a broad range of familiar basic-level categories and were controlled for average area, aspect ratio, contour variance, contrast, and luminance. On each trial, children viewed the exact image of the object they searched for. After this image was touched, it disappeared for 500ms and then reappeared among five distractors, which were either from the same broad category (Uniform displays) or a different broad category (Mixed displays). Figure 1 provides examples of Mixed displays for each experiment.

Results: Reaction times for correct trials were analyzed; participants were included in analyses if five or more trials remained in every condition after outliers were removed. In experiment 1 (N=10, M=48.4mo), children found targets faster when the distractors differed from the target in animacy (Uniform M=1978ms, Mixed M=1619ms, t(9)=3.08, p=.01). In experiment 2 (N=32, M=46.4mo), children also found targets faster when distractors differed from the target in real-world object size (Uniform M=1996ms, Mixed M=1827ms, t(31)=2.52, p=.02). Do these guidance effects exist for all possible conceptual distinctions? In experiment 3, we tested edibility (i.e., edible vs. non-edible objects), another salient conceptual dimension (e.g., Wertz & Wynn, 2014). However, we found no evidence of attentional guidance by edibility in a group of 3- and 4-year-olds (Uniform M=1828ms, Mixed M=1942ms, N=15, t(14)=-1.35, p=.20) or when we included children 5-6 years old (Uniform M=1750ms, Mixed M=1744ms, N=27, t(26)=.02, p=.98).

Discussion: Though children can identify the animacy, real-world size, and edibility of individual entities by two years of age (Mandler & Bauer, 1988; Ebeling & Gelman, 1988; Wertz & Wynn, 2014), only the broad distinctions of animacy and object size influenced visual search efficiency in 3- and 4- year olds. Within a few years, our visual system may generalize across basic level categories, constructing perceptual feature representations for animacy and object size that guide visual attention.

Long, B., Moher, M., Konkle, T., Alvarez, G.A., & Carey, S. (2015). Broad category membership guides visual attention in young children. Poster presented at the Society for Research in Child Development, March 19-21, Philadelphia, PA.