Abstract
The present study utilized a gaze-contingent adaptation of the visual world paradigm (VWP) to replicate and extend a well-known test (Biederman & Cooper, 1991) of Biederman's object recognition theory in adults and 3- to 5-year-old children. The VWP uses a priming procedure to assess the influence of a stimulus on moment-by-moment gaze of subjects following the prime. Subjects in this study were presented with a feature- or parts-deleted image, and were asked to identify that image from a four-item array containing the identical image (target), its complementary contour-deleted image (complement), a different exemplar of same object category (different exemplar), and an unrelated distractor (distractor). Gaze proportion to each of the four areas of interest (AOIs) was calculated in 24 msec bins from the onset of the array for a period of 1600 msec. Proportions were analyzed using a cluster-based divergence analysis. Adult results indicate a replication of the classic finding using eye-tracking and the VWP. Specifically, looks to the parts-based target exceeded looks to the complement early on (384 msec), indicating a lack of priming to the parts complement. Conversely, looks to the feature-based target and complement remained equivalent until much later in the trial (1032 msec), suggesting that the adults did not show a clear ability to discriminate between the two images, and indicating visual priming for the non-primed feature-based complement image. Children tested showed an overall difficulty with the feature-deleted images, in that looks to any of the four objects did not diverge until late in the trial (672 msec), at which point children appeared to show an adult-like pattern, indicating visual priming of the non-primed feature complement as well as the primed target. Children also showed an adult-like pattern when processing the parts-deleted images, but again were slower overall and showed less pronounced differences in gaze preference.
Meeting abstract presented at VSS 2017