Object-based attentional selection emerges late in visual cortical hierarchy for objects of varying perceptual strengths

Object-based attention (OBA) can - in addition to acting upon explicit object representations - act upon occluded and illusory objects. It remains unknown, however, whether or not the selection of such object representations is detectable at the same level within visual cortex. This study examined the level within visual cortex (V1-V3, LOC) at which object-based selection is observed for explicit, occluded and illusory objects. During fMRI acquisition, participants identified a target preceded by a predictive arrow cue in the double-rectangle cueing paradigm. We independently localized retinotopically-specific regions of cortex corresponding to all possible target locations to examine neural fluctuations at each level of the visual cortical hierarchy. We found, after cue onset, that activity along visual cortex was not greater for representations of cued than of uncued same object locations. In contrast, we found that activity in V3 was enhanced at retinotopic representations that correspond to uncued same than different object locations. These results, together, support attentional spreading. Additionally, when the target appeared at either the cued or uncued locations, we found higher activation in areas representing uncued same object versus cued locations. This effect emerged along the visual cortical hierarchy. Further, when the target appeared on either the cued or uncued object, we found that activation in V3 transiently increased at uncued same than different object locations. This effect was also detectable upstream in LOC. These results index attentional re-orienting between locations/objects. Effects emerged regardless of object type: explicit or completed. Thus, the gating of object information proceeds completion. Significance StatementWe investigated the level within visual cortex (V1-V3, LOC) that object-based selection is observed for explicit objects and those requiring perceptual completion. We showed that activity along visual cortex was similar for representations of locations on a cued object, which may indicate attention spreads evenly to all locations on an object marked as relevant by the cue. We also showed that activity in late visual areas was greater for representations of uncued same than different object locations, which may indicate that attention enhances the cued object. These findings support the attentional spreading account. Object selection may, thus, be instantiated by even engagement of locations within a cued, and/or suppression of locations within an uncued, object - independent of its type.