Hana Yabuki¹, Susanne Ferber¹; ¹Carnegie Mellon University

Visual working memory (VWM) allows us to hold on to representations of items that are no longer available to view. However, VWM is limited in capacity, necessitating a selection process that favours task-relevant information. Directed forgetting - i.e., the volitional removal of items from working memory, is thought of as a process that allows us to delete outdated memoranda from working memory and frees up cognitive resources. To understand the nature of these newly-available resources, we examined if they can be reallocated to incidentally encode item-context after being released. We adapted Dames and Oberauer’s (2022) WM removal paradigm. Here, participants (n = 63) were sequentially presented with a word superimposed on a background image. In the removal condition, words were followed by a ‘remember’ or a ‘forget’ cue in a clockwise pattern, with each trial consisting of 6 words. At the end of a trial, participants recalled if the probe word at test was presented in the same location during encoding. We employed two baseline conditions, set size 6 and set size 3, where all presented words were followed by a ‘remember’ cue. Directed forgetting is observed by comparing performance for trials that tested words followed by a ‘forget’ cue with words followed by a ‘remember’ cue. We replicated Dames and Oberauer’s (2022) finding that memory for ‘remember’ words was better than for ‘forget’ words, demonstrating directed forgetting in VWM. Performance for background image recognition, however, was poor, suggesting available resources do not retroactively enhance incidentally encoded context. To explore individual differences in VWM capacity and incidental encoding, further analyses will examine correlation between capacity from a change localization task and memory for background images.