Abstract
Here we investigated how ensemble statistics of a visual scene guides attentional selection and vice versa. Participants detected a target while extracting mean sizes of multiple sets of circles. On each trial, four sets of 5 differently sized circles (distinguished by color) were presented for 1000 ms, followed by a blank screen in which a target could appear on a centroid of one of the four sets in one third of the trials. After the blank screen, two color probes appeared to indicate two to-be-compared subsets for mean size comparison. When the target was presented on the blank screen, participants were asked to detect the target as soon as possible before they compared mean size of the two probed subsets. One of three different types of mean size task was assigned to each participant as a between-subject variable: participants were to judge which of two probed color-sets had larger mean size or smaller mean size, or for other participants the two tasks were intermixed and post-cued trial-by-trial. We found that participants were the most accurate in judgment of mean size when they compared the two largest color-sets in mean size, except when the task type was to choose smaller set. For the detection task, participants whose task was either to select larger mean size or randomly intermixed each trial were generally faster at detecting a target than those who selected smaller mean size. In addition, when the task was to choose larger mean size, participants’ reaction time was the fastest when the target appeared on the centroid of the set with the largest mean size. Together, these results suggest that attentional task setting influences the extraction of mean size and mean size representation guides spatial attention to the set with the largest mean size in turn without this setting.
Meeting abstract presented at VSS 2012