Abstract
How do we remember to execute a specific goal at the appropriate time (i.e., perform "prospective memory")? We can use an effortful strategy -- actively maintaining the goal in working memory -- or a reflexive strategy -- relying on external cues to trigger retrieval of the goal (McDaniel & Einstein, 2000; Reynolds, West, & Braver, 2009). The flexible choice of strategy involves a tradeoff between effort and performance, and depends on the availability of cognitive resources and the likelihood of goal retrieval. We designed a fMRI study to track, on a trial-to-trial basis, the use of working memory versus episodic memory to achieve a delayed goal. Each trial began with a target picture, followed by a variable-length sequence of 2-sec memory probes, each containing two pictures and a string of letters. Participants were required to make repeated lexical judgments about the letter strings until the picture target reappeared (between 2 and 42 sec after its introduction). We varied the difficulty of the lexical judgments and the degree of proactive interference associated with the target. We reasoned that when a trial involved lexical judgments that were more demanding, participants would be less likely to maintain the picture target in working memory, relying instead on retrieval from episodic memory. On other trials, high proactive interference would interfere with episodic memory retrieval, thereby biasing participants to use working memory to maintain the target. Multivariate pattern analysis of fMRI was used to identify target-related activity during each trial. Prospective memory success was associated with elevated estimates of target activation at the moment it reappeared (which could indicate either working memory or episodic memory use). Critically, target activations prior to its reappearance (indicative of working memory use) were more predictive of success on trials designed to bias the use of a working memory strategy.
Meeting abstract presented at VSS 2014