Abstract
Shifts of visuospatial selective attention are limited in spatial resolution (i.e., the required spacing surrounding a target for successful selection; Intriligator & Cavanagh, 2001). Previous research on attention resolution has focused on endogenous selection, but exogenous resolution has been unexplored. We measured the resolution of exogenous shifts of visuospatial attention by quantifying the minimum spacing needed for individuals to isolate and select a peripheral target among nearby distractors. In three experiments, participants viewed a circular array of equally spaced, luminance-matched colored disks (peripheral stimulus diameters of 1°, 0.75°, or 0.5°, respectively) at 10° eccentricity on a median gray background with a single white RSVP stream at fixation. Subjects monitored RSVP items for one or more target digits and responded via button press. Simultaneously, on each trial a black dot briefly (60 msec) appeared (among the peripheral-colored disks) which exogenously captured attention. After each trial, participants selected via mouse-click the color (from several choices) corresponding to the location nearest to which the black dot had appeared. Participants in Experiment 2 (M = 49.5%) performed significantly better than participants in Experiment 1 (M = 37.4%) and Experiment 3 (M = 32.9%) on the color selection task. Responses were compared to an ideal observer model to determine the attentional window size each subject deployed in response to the exogenous cue. This window size was then used to quantify the minimum stimulus spacing needed for performance to reach 75% correct. Minimum spacing estimates were not significantly different between Experiment 1 (M = 1.076°), Experiment 2 (M = 1.051°), and Experiment 3 (M = 1.241°) or within-subject (across the visual field). We conclude that increasing the spacing between adjacent peripheral stimuli improves the resolution of attention (via reduced perceptual crowding) to a point, but is detrimental when peripheral items become too small to resolve.