Abstract
Visual attention has been conceptualized as being a mechanism of selecting a target location or object (e.g. Intriligator & Cavanagh, 2001), and as a medium that can be distributed across a span of the visual field (e.g. Eriksen & Hoffman, 1974). These two approaches to understanding attention, however, have typically been investigated separately. We asked whether the effects of selecting a location in the visual field would be mediated by the way attention was distributed across possible target locations. To answer this question, subjects performed a letter identification task while orienting their attention endogenously to one of several squares on a computer screen. In Experiment 1, eight squares were presented in a line, four on either side of fixation, increasing in size toward the periphery to account for cortical magnification. The line of squares could be oriented on the vertical or horizontal meridian, or at intercardinal locations halfway between the cardinal meridians. If invalidly cued (30% of trials), the target would always appear on the opposite side of fixation, in order to require subjects to distribute their attention across the entire display to the extent needed to maintain high levels of performance. In Experiment 2, the same stimulus configuration was used, but rather than a fixation cross at center, there remained the possibility that a target might appear in a centrally-located box at fixation. Results showed that maintaining attention at fixation, as in Experiment 2, facilitated performance closer to fixation to a greater extent than in Experiment 1, revealing an attentional gradient anchored at fixation. Across the two experiments, attentional facilitation was anisotropic across the visual field, especially in the right visual field on the horizontal meridian. Our findings reveal an intriguing interplay between distribution of attention across possible target locations and the ability to attentionally select a target location.