Fast Access Categorization of Familiar Shapes Can Be Independent from Perception

Abstract

Visual perception has typically been explained using a feed forward processing model. Several studies indicate that categorization of familiar objects can take place using this processing method. I investigated whether categorization of familiar versus novel silhouettes was slowed when another familiar object is potentially present and not perceived. Through 80 trials, subjects were shown two silhouettes, one familiar and one novel, with the task of finding the familiar silhouette. Familiar silhouettes represented everyday objects, while novel were never seen before. Half of the novel silhouettes contained familiar shapes suggested on the outside edges but not perceived. The average saccade toward the familiar shape during the experimental silhouette condition was significantly slower in both Experiment 1 (281 ms vs. 267 ms, p < .02) and Experiment 2 (217 vs. 208 ms, p < .01). This study showed that although fast categorization can take place, there must be mechanisms of feedback for perception to ultimately occur.