Reexamining Object-Based Visual Attention: Understanding the Nature of Direction-Dependent Attention Shifts

Abstract

Attentional selection is a process by which relevant sensory stimuli are afforded enhanced priority for processing over and above irrelevant sensory stimuli. Object-based attention is a form of selection that leads to preferential processing of visual information contained in/on an attended object versus an unattended object. Observers typically exhibit enhanced performance to a target at an invalidly cued same object location compared to a different-object location, which results in a same object advantage as measured by the response time (RT) difference between these two target locations. A growing body of research has shown that object-based effects are small, inconsistent, and unreliable. Nevertheless, previous work showed larger same object advantages for horizontally oriented rectangles than vertically oriented rectangles (Pilz, Roggeveen, Creighton, Bennett, & Sekular, 2012). To explain this effect, it was postulated that attention may be more efficiently allocated along the horizontal visual field midline (i.e., meridian) than the vertical midline. Here, our goals are to (1) disentangle the confound between shift direction, object orientation, and object selection/competition and to systematically compare a new metric of object-based attention, the Shift Direction Anisotropy (SDA), to the standard measure of object-based attention, the same object advantage, (2) determine whether the SDA depends upon meridian crossings of object boundaries, target locations, or both, (3) causally implicate the meridians in the emergence of the SDA by examining its susceptibility to perceptual enhancements of the meridians, and (4) characterize the neural correlates of the SDA. In the end, we demonstrate that the SDA is more larger, consistent, and reliable than the same object advantage, that the SDA is driven by meridian crossings of the invalid target locations, that the SDA is malleable and susceptible to strong perceptual manipulations of the horizontal meridian, and that functional cue-related and target-related neuroimaging data mimic the behavioral SDA. In sum, this work (1) introduces a novel method for investigating anisotropic shifts of object-based attention that have been previously observed in the literature (2) provides the foundation for a comprehensive investigation into the effects of the visual field meridians on real-world object-based attentional selection, and (3) directly challenges and updates current theories of object-based attention to account for visual field and neuroanatomical constraints.