Abstract
Object-based attention yields a performance advantage for targets on the same object compared to targets on different objects. Although most published reports find evidence of a same-object advantage (SA), certain situations evoke a same-object cost (Davis & Holmes, 2005). Pilz et al. (2012) reported that most individuals do not show SA, and SA is more prevalent for rectangular objects oriented horizontally than vertically; consistent with attention being more efficiently allocated along the horizontal meridian. To explore these object-based effects when explicitly controlling for shifts of attention across either the horizontal or vertical meridian, we reanalyzed published data from four experiments (Pilz et al., 2012; Greenberg, 2009). In each experiment, rectangle orientation (horizontal vs. vertical) was a factor, allowing computation of SA while controlling for attention shift direction. We controlled for shifts across meridians by subtracting the different-object RTs for one object orientation from the same-object RTs for the other object orientation, thus ensuring that subjects' attention focus never crossed a specific meridian. Results showed that (1) controlling for meridian shifts failed to produce a same-object cost at the group level, (2) shifts within one side of the vertical meridian produced larger SA than for the horizontal meridian, and (3) a smaller proportion of individual subjects showed same-object costs when controlling for meridian shifts. Furthermore, when the target was part of the object (cf. Watson & Kramer, 1999), there was no difference in SA between our meridian analyses, unlike when the target was placed on the object. We conclude that controlling for attention shifts across meridians provides an important perspective on SA. Object-based effects are stronger when attention shifts are confined to one side of the vertical meridian (vs. horizontal). Examining orientation differences confounds meridian shifts (and therefore brain hemisphere representations), which may explain inconsistent SA observations in the literature.
Meeting abstract presented at VSS 2014