Abstract
Recently there has been considerable interest and debate surrounding the capacity of visual working memory. Estimates of working memory have declined steadily from the “magic number” seven, to four, and even recently to very nearly one item. Though the number of items able to be stored in visual working memory is arguably few, there has been no controversy regarding the belief that visual working memory consists of a single store.
The present study, however, was inspired by a growing number of findings indicating that the brain can maintain multiple states simultaneously and switch between these states. For example, Dobbins and colleagues (2004) revealed that repetition priming is contingent upon how the observer intends to respond: whereas repetition priming for an object exists under one set of instructions, it is eliminated when adopting a second set of instructions, but returns when the observer switches back to the original set of instructions.
Does the same hold true for the capacity of visual working memory? To answer this question, we required observers to remember two working memory arrays: one array contained shapes and the other contained colors. The observers' task was to detect a change in the probed display. Importantly, they did not know which array would be probed until the test array appeared. The data show that performance on the probed array was not affected by the number of items in the non-probed array, i.e., the capacity of working memory was able to be multiplied when observers presumably maintained a separate ‘page’ for each display. This outcome requires a reconsideration of the capacity of visual working memory as not a fixed value of a single store; rather, visual working memory can be expanded by the maintenance of multiple stores.
The authors thank Jolanda Roelofsen for her help with data collection. Funding for this study was provided by the Canadian Institutes of Heath Research and the Nederlandse Organisatie voor Wetenschappelijk Onderzoek.