Abstract
Our goal is to better understand object-based attention using the spatial filtering paradigm. In our version of spatial filtering (Palmer & Moore, submitted), observers are asked to detect stimuli at a cued location (targets), while ignoring otherwise identical stimuli that are presented at nearby locations (foils). Contrast of targets and foils are varied in order to obtain psychometric functions for both the target and the foil at various spatial separations. The foil psychometric function provides a measure of the extent to which stimuli in uncued locations are processed despite efforts to ignore them. Results from this paradigm show large effects of separation on the asymptote of the foil psychometric function and little or no effect on the threshold. These results rule out contrast gain as the primary mechanism of selection, and instead are consistent with an all-or-none mixture as the selection mechanism. We used this paradigm to measure an effect of perceptual structure on spatial filtering. We presented an outline rectangle in the visual field near where targets and foils appeared. The rectangle provided no specific information about the target location, but instead provided a visual context on which the spatial filtering paradigm was superimposed. Results showed an effect of perceptual structure on spatial filtering. Specifically, selection of a target location across a boundary from a foil location was substantially better than when they were not separated by a boundary. Additional experiments investigate the extent to which spatial filtering in this context can be accounted for by a contrast gain or an all-or-none mixture model.
Supported by NIH grant MH067793 to C. M. Moore.