Abstract
We tested adults and children aged 7 and 14 on the ability to integrate contour elements across variations in the collinearity of the target elements and in their spatial proximity. Participants were asked to find the 14 Gabors forming an egg-shaped contour among randomly positioned background Gabors. Across trials, the density of background noise Gabors was varied according to a staircase procedure to determine a threshold for each combination of collinearity and spatial proximity. Thresholds were expressed as the relative density of the background noise elements to target elements (Δ).
When the collinearity of the target Gabors was high, the thresholds of adults (n = 24 in each of Experiments 1 and 2) were largely independent of spatial proximity and varied only with Δ. It was only when collinearity was less reliable because the orientation of the elements was randomly jittered that spatial proximity began to influence adults' thresholds. These patterns correspond well to the probability that real-world contours compose a single object: collinear elements are more likely to reflect parts of a real object and adults integrate them easily regardless of the proximity among those collinear elements. The results from 7- and 14-year-olds (n=24 per age group) demonstrate a gradual improvement of contour integration throughout childhood and the slow development of sensitivity to the statistics of natural scenes. Unlike adults, integration in children at both ages was limited by spatial proximity regardless of collinearity and one strong cue did not compensate for the other. Only after age 14 did collinearity, the most reliable cue, come to compensate efficiently for spatial proximity.