Abstract
An emerging framework suggests that visual working memory (WM) representations rely on the same representational resources as those used to process external visual input. Kiyonaga and Egner (2014) provided support for this claim by demonstrating with a modified WM Stroop task that an irrelevant color word held in WM produces the same Stroop interference patterns on a perceptual color target as that seen in a classic perceptual Stroop task. However, there is evidence that perceptual and WM representations differ in terms of their spatial representation. Specifically, unlike visually perceived stimuli, the representation of information in WM might be spatially global rather than tied to any specific retinotopic location (Ester, Serences, & Awh, 2009). To test the spatial specificity of WM representations, we compared a classic perceptual Stroop task with a WM version in which the color word and color patch appeared either in the same or a different spatial location. In Experiment 1, for the perceptual version, we found a spatial effect such that spatial separation eliminated Stroop interference. However, in the WM Stroop task, robust Stroop interference was demonstrated in both the spatially overlapping and spatially separated conditions. In Experiment 2, we ensured location of the color word was encoded into WM by only testing memory for the color word's location. At test, the color word either appeared in the same spatial location or at a spatially displaced location—the extent of which was modified throughout the task to ensure motivated encoding. Replicating the results of Experiment 1, spatial separation eliminated Stroop interference for the perceptual task, but was present for both spatially overlapping and spatially separated conditions of the WM Stroop task. These experiments support the conclusion that at least in terms of spatial information, perceptual and WM representations do not rely on the same neural machinery.
Meeting abstract presented at VSS 2016