Abstract
Mechanisms of feature-based attention underlie facilitated processing of the attended feature throughout the visual field. We have previously shown this holds with collinearity: we demonstrated that perception of collinear-defined loops was better when attending to other collinear-defined loops rather than attending to motion-defined loops. Our results could provide evidence for high-level feature-based attention. Alternatively, it is possible this advantage was simply due to a low-level effect of symmetry. That is, presenting collinear-defined loops together with other collinear-defined loops created a more symmetrical display than when presented together with motion-defined loops. Here we investigated symmetry as a factor in explaining the effect of feature-based attention. To disrupt symmetrical processing we presented S-shapes together with loops. The experiment used a dual task paradigm, that included two concurrent rapid serial presentation streams of scattered gabors appearing in the left and right visual hemifield. The primary task, indicated by a fixation arrow, presented two S-shapes that were either (a) collinear-defined or (b) motion-defined. Coinciding with one of the two S-shapes was the appearance of the secondary collinear-defined loop. Perceptual thresholds were determined in a two-alternative-forced-interval fashion by manipulating the amount of orientational/rotational noise of the gabors. The results replicated our initial experiment in that attending to collinear-defined objects facilitated the perception of collinearity on the unattended side. These results suggest that symmetry cannot explain the observed effects of feature-based attention. Furthermore, what appears to be important is not the global structure of the object but a specific feature - the collinearity of the object's contour.