Cognitive Control, Not Time, Determines the Status of Items in Working Memory.

Jacqueline M Fulvio,Bradley R Postle

doi:10.5334/joc.98

Abstract

Despite the fact that multiple items can be held in working memory (WM), it is often the case that only one of these is relevant for guiding in-the-moment behavior. Therefore, understanding how priority is established and controlled in WM is an important problem. Data from Rose et al. (2016) have provided evidence that although neuroimaging evidence for an active trace of an “unprioritized memory item” (UMI) held in WM drops to baseline levels, evidence for its retention in WM can be “reactivated” by a single pulse of transcranial magnetic stimulation (TMS). Critically, this TMS-reactivation effect was specific to the first delay period of a dual serial retrocue (DSR) task, when the UMI could be needed for the trial’s second memory probe, and was not observed during the second delay period, when the uncued item was no longer needed (i.e., when it is an “irrelevant memory item” [IMI]). A problem for the interpretation of these results, however, is that the status of the UMI/IMI was confounded with time spent in WM, as well as with the number of intervening cognitive operations. Here, we report data from a follow-up study designed to replicate the findings Rose et al. (2016) and to add a condition that unconfounds time-since-sample-presentation and UMI/IMI status. The results indicate that the TMS-reactivation effect is, indeed, an index of status in WM (UMI vs. IMI), and not a mere consequence of time elapsed since sample presentation.

Highlights

An important question in working memory (WM) research is how we maintain multiple items in WM and dynamically prioritize them according to task demands (Myers et al, 2017)
We investigated the impact of delay-period single pulses of transcranial magnetic stimulation (spTMS) on behavioral performance implemented in two working memory tasks, a dual serial retrocuing (DSR) task like that used in Rose et al (2016) and a single retrocue task in which subjects are cued and tested on the memory set only once
There were three conditions of interest concerning the impact of spTMS during (1) the delay period of the DSR task when the uncued item had the status of unprioritized memory item” (UMI); (2) the delay period of the DSR task when the uncued item had the status of IMI; and (3) the delay period of the single retrocue task, when the uncued item had the status of IMI

Summary

Introduction

An important question in working memory (WM) research is how we maintain multiple items in WM and dynamically prioritize them according to task demands (Myers et al, 2017). The core finding, from functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) studies, has been that multivariate evidence for an active representation of a sample item drops to baseline when it takes on the status of UMI (Lewis-Peacock et al, 2012; LaRocque et al, 2013; LaRocque et al, 2017). This suggests that the UMI may transition to a representational state that is different from when it is attended. This suggests that the UMI may transition to a representational state that is different from when it is attended. (Note that more recently there have been reports of evidence for an active neural representation of the UMI [e.g., van Loon et al, 2018; Yu & Postle, 2018; Christophel et al, 2018; Rademaker et al, 2019; Wan et al, preregistered], but this development in the literature is not directly relevant for the present study.)

Methods

Results

Conclusion