Abstract
Humans have the capacity to selectively attend to more than one item in the visual field. Whether this is accomplished by a single scalable attentional window, grouping targets into a virtual polygon, serial monitoring of targets, or divided multiple foci of attention has remained a controversy. Recent neurophysiological results have demonstrated divided focus of attention in humans and monkeys (Neibergall et al., 2011; Stoermer et al., 2013), but evidence for endogenously selected multi-focal attention that can differentiate between various mechanisms has been elusive. One strategy is to exploit the close linkage between ocular fixations and spatial selection (Goffart, Hafed & Krauzlis, 2012; Fehd & Seiffert, 2008). We used a novel attentional task paradigm where subjects selectively attended multiple static target locations while ignoring distractor locations within a geometric array. The arrays consisted of letters that rapidly changed identity at a rate of 5 or 10 Hertz. The subject's task was to detect the occurence of probe numeral that was briefly displayed (single frame) at one of the target locations. Importantly, in contrast to standard attentional tasks, subjects were free to select their own preferred fixation position. We found that the mean preferred fixation positions were close to the centroid defined over a group of discrete circular windows encompassing target locations only, consistent with divided multi-focal attention. Fixation positions were not consistent with selection based on a single attentional window or a virtual polygon defined by target locations. Moreover, analysis of differences in the detectability of probes in configurations with 3 or 5 target items were incompatible with serial covert monitoring of targets. Our results are consistent with the view that ocular fixation represents the equilibrium point of the spatial distribution of discrete multiple foci of attention and suggest a more nuanced definition of covert versus overt attention.
