Abstract
Adults can see bounded figures even when local image information fails to provide cues to specify their edges (e.g., Ginsburg, 1975). Such contour interpolation leads to the perception of subjective contours — edges perceived in the absence of any physically present discontinuities (e.g., Kanizsa, 1955). We examined the development of sensitivity to shape formed by subjective contours and the effect thereon of support ratio (the ratio of the physically specified contours to the total edge length).
Children aged 6, 9, and 12 years and adults (n = 20 per group) performed a shape discrimination task. Sensitivity to shape formed by luminance-defined contours was compared to shape formed by subjective contours with high support ratio (interpolation of contours between the inducers was over a small distance relative to the size of the inducers) or low support ratio (interpolation of contours was over a larger distance).
Results reveal a longer developmental trajectory for sensitivity to shape formed by subjective contours compared to shape formed by luminance-defined contours: only by 12 years of age were children as sensitive as adults to subjective contours (pp[[gt]].1). The protracted development of sensitivity to subjective contours is consistent with evidence for delayed development of feedback connections from V2 to V1 (Burkhalter, 1993) known to be important for the perception of subjective contours (e.g., Ffytche & Zeki, 1996). As in adults, sensitivity to subjective contours was better with higher support ratio by 9 years of age (psp[[gt]].1). The results suggest that, over development, support ratio becomes a reliable predictor for interpolation, so that contours that are more likely to reflect real objects‘ contours (i.e., highly supported contours) are more easily interpolated.
Supported by the Canadian Institutes of Health Research, grant # MOP-36430 and the James S. McDonnell Foundation