Abstract
The storage capacity of visual working memory (VWM) is strongly correlated with broad measures of cognitive abilities, but the nature of capacity limits has been the subject of considerable controversy. Some researchers have proposed that VWM stores a limited set of discrete, fixed-resolution representations, whereas others have proposed that VWM consists of a pool of resources that can be allocated flexibly to provide either a small number of high-resolution representations or a large number of low-resolution representations. To distinguish between these possibilities, we tested whether observers can, depending on task demands, store either a larger number of coarse-grained representations or a smaller number of fine-grained representations. The flexible resource hypothesis predicts that observers should be able to strategically trade off resolution for capacity. In contrast, the fixed resolution hypothesis predicts that VWM resolution and capacity are not under strategic control.
We used a color recall task in which observers attempted to retain several colors in VWM over a 1-s retention interval and then recalled one of them by clicking on a color wheel. Task demands for VWM precision were manipulated in different ways in three experiments. For example, high precision was necessary when the color wheel contained 180 continuously varying colors, whereas low precision was sufficient when the color wheel was divided into a small number of discrete color wedges. In addition, differential payoffs were used to reward either high capacity or high resolution. Across the three experiments, we consistently found that VWM resolution and capacity remained constant across conditions, supporting the fixed resolution slot hypothesis. In sharp contrast, observers could trade off capacity and resolution when iconic memory was tested by eliminating the delay between sample and test. Thus, working memory consists of a limited set of fixed-resolution representations, whereas iconic memory consists of a flexible resource.
This research was made possible by grants R01 MH076226 (S.J.L.) & 1F32MH090719 (W.Z.) from the National Institute of Mental Health.