Object-based attention improves memory fidelity for unattended same-object stimulus, but at a cost to the attended stimulus

Accounts of object-based attention (OBA) suggest that when attention is directed to one part of an object, a benefit in processing is incurred for other parts of that same object. However, studies of OBA typically do not simultaneously measure representations of items at the cued and uncued locations. Here we adapt the classic Egly et al. (1994) paradigm to measure memory representations with continuous reproduction. In Experiment 1, subjects were cued to a location on one of two visual objects. Following the cue, two colors were presented, one in the cued location and the other in an uncued location on either the same or different object. After a delay, subjects then reported one of the colors (with 80% cue validity) using a continuous report color wheel. As predicted by OBA accounts, on invalid trials, we find the uncued same-object location had a more accurate representation than the uncued different-object location t(45) = -2.89 , p< 0.01. However, on valid trials we found a new kind of interference: a cost to the representation of the cued item when the uncued item was in the same-object compared to when the uncued item was in the different-object t(45)=2.558, p<0.05. Finally, we found that the two presented items reliably interacted: the cued item was repelled from the uncued item, whereas the uncued item was attracted toward the cued item. In Experiment 2 we replicated these effects using a continuous shape space to generate stimuli. Overall, we find evidence for a novel interference effect in OBA: representations of the cued object are less precise when another item is presented on the same object. We also demonstrate that OBA can be reliably measured using continuous reproduction, opening avenues for further research on the effects of OBA on memory representations.