INFLUENCE OF TASK AND STRATEGY ON THE NEURAL AND BEHAVIORAL CORRELATES OF THE FOCUS OF ATTENTION

Abstract

Working memory (WM) is often described as a mental workspace where information can be maintained and manipulated in the service of ongoing cognition. Theoretical accounts describe the focus of attention as a state within working memory where a limited number of items can be briefly maintained in a heightened status of awareness. Ongoing debate and conflicting empirical evidence surrounds the capacity and characteristics of the focus of attention. Substantial recency effects are reported in a group of WM studies, and these recency effects are interpreted as a marker of the focus of attention (e.g., Nee & Jonides, 2008; Oztekin, Davachi, & McElree, 2010). The present work considers whether these findings are specific to parameters of these particular studies or whether they generalize across a broader range of tasks. An initial behavioral experiment tested performance across two tasks (judgment of recency and judgment of primacy), two information types (verbal and spatial), and two self-reported strategies (maintenance-based and retrieval-based). Central analyses averaged trials by the serial position of the correct item, and compared the accuracy and speed of retrieval of trials in different serial positions. Results showed evidence of both recency effects and primacy effects in all four types of task (verbal judgment of recency, verbal judgment of primacy, spatial judgment of recency, and spatial judgment of primacy). Moreover, a significant task by effect-type interaction showed that the size of recency and primacy effects shifted with the demands of the task (e.g., larger recency effects in judgment of recency than in judgment of primacy). Some similarities and some differences were found between verbal and spatial domains, while no differences were found across self-reported strategy. A subsequent fMRI experiment examined the neural correlates of verbal judgment of recency and primacy. Again, behavioral results showed a task by effect-type interaction where there was a larger recency effect in judgment of recency and a larger primacy effect in judgment of primacy. FMRI results showed no distinct correlates of a recency effect. In other words, , contrasts comparing fMRI signal during retrieval of recency item trials and middle item trials did not reveal above threshold clusters of activation. In contrast, neural correlates of primacy were found in frontal lobe brain regions (BA 4, 6, 32) associated with active maintenance of information. Moreover, the precise neural correlates of primacy were task-specific. In sum, two experiments demonstrate that the behavioral and neural signatures of WM, specifically related to primacy and recency effects, are dependent on task-demands. Accounts of the architecture of WM should address these observations, which inform how competing claims are supported across studies of WM.