Abstract
Previous studies have demonstrated that observers can select and track objects independently in the left and right visual field, as if separate attentional systems were engaged in each hemifield (Alvarez & Cavanagh, 2005). This previous work required observers to track objects that were always contained completely within one half of the visual field. However, in real world viewing conditions, tracked objects often move freely from one hemifield to another. We tested how attention selects and tracks objects as they move across hemifields by comparing performance when tracked items shift from one quadrant to another, either within one hemifield, or crossing between hemifields. On each trial 8 items were presented in diagonally opposite quadrants (e.g. 4 in the top left and 4 in the bottom right), and two items in each quadrant were highlighted as targets for tracking (4 targets total). All of the items began to move, then at one point during the trial all of the items shifted to an adjacent quadrant, either staying within the same hemifield (vertical shift), or crossing over (horizontal shift). At the end of the trial, all of the items stopped, and observers had to click on the target items. Observers tracked targets less accurately when the targets crossed hemifields (70%) than when they remained within a hemifield (76%, t(9)=2.82, p<.05). The relative distance the target items shifted was equal, whether the shift occurred within or across hemifields. Thus, these results suggest that the critical factor is whether the targets remained within one half of the visual field, or shifted across the left and right hemifields. The difficulty of tracking objects that shift across the left and right hemifields could reflect a cost in “handing off” the tracking of an object between the attentional systems engaged in the left and right hemifields.