Abstract
Visual Working Memory (VWM) representations might require active maintenance, but research has focused predominantly on storage as opposed to the updating and manipulation of these representations. For example, VWM might maintain three unique features bound to three unique positions with little or no cost (Luck & Vogel, 1997; but see Bays & Husain, 2008), but it remains relatively unknown how these VWM representations may or may not suffer under conditions of manipulating and updating. Here, we investigated the costs of manipulation in VWM using a variant of a shell game. Participants were presented with a memory display consisting of 2, 3, or 4 colored circles. These colors then disappeared leaving only uniform outlines of these objects. Pairs of objects then underwent smooth motion and switched positions. The number of switches varied from 0 (static) to 4. After switches, one object was filled and subjects judged whether it was the expected or an unexpected color. Results suggest that there is little or no cost to updating color information for 2 stored representations. However, while participants were accurate at maintaining 3 and 4 individual static colors (consistent with classic results, e.g., Luck & Vogel, 1997), they showed significant cost when updating of these color-position bindings was required. Consistent with some previous tests of updating in non-visual memory domains (e.g., counters, Garavan, 1998; Feigenson & Yamaguchi, 2009), our experiments reveal somewhat independent limits for the storage of static information and the active manipulation of information in VWM.
Meeting abstract presented at VSS 2012