Abstract
Object-based attention (OBA) preferentially enhances visual information within the boundaries of attended versus unattended objects. The double-rectangle cueing paradigm (Egly et al., 1994) typically exhibits enhanced performance to targets at a same-object versus different-object location (~13ms), termed the “same-object advantage”. Mounting evidence suggests this effect is inconsistent and small in magnitude: several studies have failed to show a same-object advantage or have even found a same-object cost. Moreover, Pilz and colleagues (2012) demonstrated that these effects vary by object orientation, with few participants (15%) exhibiting significant effects. This inconsistency ultimately encourages questions regarding the legitimacy of OBA. In the double-rectangle cueing paradigm, confounds between shift direction, object orientation, and object selection may drive these inconsistencies. We developed a paradigm that eliminated these confounds by permitting attention shifts across the visual field meridians to be contained within the boundaries of a single ‘L’-shaped object. Following a peripheral cue, participants detected a target that appeared at the cued location or at one of two noncued locations equidistant from the cue – the horizontal (“invalid-horizontal”) or the vertical (“invalid-vertical”) locations. At the group level, RTs were significantly faster at the invalid-horizontal versus invalid-vertical location, resulting in the “shift direction anisotropy” (SDA; ~78ms). Relative to the proportion of individuals who exhibited a same-object advantage reported by Pilz and colleagues (2012), a significantly larger proportion of participants exhibited an SDA (65%). A within-subjects experiment revealed larger and more prevalent effects for the SDA versus the same-object advantage. These findings demonstrate that the confounds described above might have caused past inconsistent OBA results and that the SDA may be a more reliable measure of OBA than the traditional same-object advantage. Thus, stable and large magnitude effects of OBA selection do exist when examined from a perspective that ameliorates confounding factors.