EFFICIENT OR INEFFICIENT PROCESSING OF SHADOWS? A CHANGE BLINDNESS STUDY
A change detection experiment is reported in which we addressed whether cast shadows of simple objects engage attentional processing using a flicker one-shot technique. On each trial, participants were shown two images separated by a blank frame, and asked to report whether a change had occurred in the second image or not. Images consisted of 12 synthetic objects casting shadows spatially arranged in a fan-shaped configuration. A change was present in only 50% of trials and consisted of one of the objects casting a shadow with an incongruent shape, a shadow with a different lighting direction, or both. Two conditions were created: a real shadow 3D condition, and a 2D condition (obtained through reversed contrast polarity). Change detection performance was overall poor, although higher in the real shadow condition than in the negative image condition. We conclude that cast shadows are processed inefficiently, but significantly better than 2D shapes.