Abstract
A salience map is a topographical map where only salience (importance) is recorded at each spatial location without any record the features that produced the salience (Koch and Ullman, 1985). Substance indifference is a defining feature of salience maps. Although used originally to order the action-priority of locations, Sperling & Lu (1995) and Gan, Sun, & Sperling (2021) showed that that third-order motion direction and distance judgments between two equal size items in the frontal plane are computed on a salience map. Here, we demonstrate that centroid judgments also are based on the information recorded in a salience map. We presented subjects with stimuli containing 32 stimulus items for 300 msec immediately followed by a 50-msec blank field and a 100-msec mask. Three experienced subjects estimated the centroid (“center of gravity”) of all stimulus items by moving and clicking a mouse cursor. Each subject completed 400 trials with complete feedback after each trial. The critical variable was the composition of the stimulus items: 32 items had either the same feature or 2, 4, or 8 different features. Items varied in color, shape, and luminance (<,=,>) relative to the background. Eight different compositions of stimulus items were tested in a mixed-list design. Results: For all subjects, centroid judgment accuracy was statistically equivalent for all 8 compositions. Being of identical composition offered zero advantage for judged centroid accuracy over being heterogeneously different. Conclusion: Centroid judgments are substance indifferent and are made on the basis of a salience map.