Abstract
The current study explores two issues about the relationship between working memory (WM) and attention. First is whether information retrieved from long-term memory (LTM) into WM guides attention. The second issue concerns the level of abstraction at which WM representations guide attention. The method is adapted from Downing (2000, Psychological Science, 11:6, 467–473). At the start of each trial subjects are given an item to hold in WM. Then two cue objects are simultaneously flashed on the screen, one of which matches the memory item. A square is then presented at the location of one of the flashed items and subjects respond as to whether it has a gap at the top or the bottom. After response to the square, a test item is presented and subjects indicate if it is the same as the memory item. If WM guides spatial attention, then RTs to the probe should be faster when they occur in the same location as a cue matching the memory item. In short-term memory (STM) conditions a memory item itself was presented at the start of each trial. In LTM conditions the basic-level name of a memory item was presented and subjects had to retrieve the item from LTM from a set of 6 objects they memorized earlier. For both STM and LTM conditions, subjects completed one block of trials where the two cue items flashed on the screen were different exemplars from the same category and another in which they were from different categories. Results from the STM/Different category condition replicate Downing (2000) in that most subjects were faster to respond to the probe when it appeared at the location where a cue matching the memory item was flashed, indicative of an attentional shift to that location. In all other conditions there was no evidence that the contents of WM guided attention in this way. The results suggest 1) items recalled from LTM do not guide attention in the same way as items held in short-term WM and 2) WM does not guide attention at a fine level of detail.