How repetition hinders priority switching in visual working memory states: An EEG study

Abstract

AbstractWorking Memory (WM) prioritizes items as per the requirement of a top-down goal. However, bottom-up attention might interfere with this prioritization as it can act as a distractor to goal-driven tasks. Bottom-up attention induced by saliency and relational account, like the perceptual similarity between items, has been found to facilitate and attentionally prioritize while recalling items during short-term working memory. The top-down goal tries to maintain and selectively recall items in differently prioritized WM states, a hierarchical way beneficial in recalling information. However, there is a genuine knowledge gap about whether repeated items, a relational property of stimuli, can acquire a prioritized WM state by default and act as a distractor that hinders recalling not-repeated items even if a matching probe prioritizes them. To address this gap, we designed a novel Visual Working Memory based EEG study where a memory array comprising of repeated and non-repeated items was introduced, and a probe item was presented after a delay period; participants responded by matching the probe to items that might be present or absent in the memory array. We found significantly slower reaction times and comparatively poor accuracy for recalling not-repeated items, which suggests that not-repeated items are not prioritized by default. Using EEG-based spectral perturbation-based analysis, we identified specific differences for sensor clusters and increased power of alpha, theta, and the beta band as the neural correlate of probe matching for not-repeated vs. repeated conditions. Comparatively higher mid-frontal theta power for not-repeated probe trials than repeated probe trials indicating stronger WM control for the matching probe. Also, significantly higher parietal alpha power for the not-repeated probe matching condition indicates a role of active inhibition of repeated item representations prioritized by default and hinders recalling accurately deprioritized items. These results imply that despite both items being encoded from the memory array in a limited time, repeated items possibly hold on to the default prioritized WM state. In contrast, not-repeated items acquire a secondary WM state. Higher frontal theta and parietal alpha power, along with behavioural results, demonstrate switching of WM states to prioritize hindered WM state corresponding to not-repeated probe. In summary, this opens up avenues for further investigations of mechanisms of priority switching from default prioritized items to goal-driven relevant items in short-term visual working memory.