Abstract
Visual working memory (VWM) is our mental chalkboard: We can hold visual images actively in mind, and then, when we are done, erase them. But are the items we previously held in VWM ever truly erased, or do they, like a real chalkboard, leave behind smudges that interfere with subsequent memories? Some work has claimed to find effects of previously held items in VWM tasks, but these results are controversial because they could derive from participants making use of not only VWM, but also long-term memory (VLTM). Here we examined proactive interference in VWM for colors by using a paradigm designed to isolate the influence of VWM vs. VLTM traces. Participants (N=50) were shown 4 outlines of circles, which were briefly filled in with colors. Then the circles rotated such that each outline was now one spot clockwise or counterclockwise from where it began, and participants reported what color would now be in a particular cued location by choosing from a set of 7 options. Two of the answer choices always corresponded to colors that had (1) been originally shown in the tested location on the previous trial (proactive-before-rotation), and (2) ended up in the tested location on the previous trial in VWM (proactive-rotated). Because participants never saw the items in their updated locations, and had to actively manipulate them, we reasoned that any interference from the previous trials' updated locations must derive from VWM, not VLTM. We found the largest interference effects were those with a VWM locus: participants incorrectly picked the proactive-rotated color more than a color that had not appeared on the previous or current trial (6.5% vs. 5.8%; p=0.04). We replicated this effect again in a new sample (N=40; p=0.025). This suggests that VWM is not truly erased between trials, even in color working memory.
Meeting abstract presented at VSS 2017