Sahcan Özdemir¹, Eren Günseli², Daniel Schneider¹; ¹Leibniz Research Centre for Working Environment and Human Factors, ²Sabanci University

It has been proposed that action control mechanisms can influence gating of visual working memory (WM) from task-irrelevant sensory input. Further, we hypothesized that a relevant action pattern could trigger WM updating through this mechanism, facilitating the interference of irrelevant sensory input on WM. To examine this idea, we conducted a preregistered (OSF) EEG experiment using a delayed-match-to-sample task. Participants memorized the color of a target stimulus and, after a retention period, identified its color on a color wheel. The target’s shape indicated which hand they should use to adjust the wheel. During the retention phase, participants performed a secondary task requiring a response with either the right or left hand, which either matched or mismatched the hand used in the primary task. This secondary task introduced two types of interference: motor interference or visuomotor interference, with the latter involving a task-irrelevant color of the cue. First, EEG analysis revealed mu/beta suppression during WM encoding, indicating simultaneous preparation of action plans with visual encoding (H1). Second, visuomotor interference caused greater disruption to target representation compared to motor interference, as reflected in WM task accuracy (H2). Finally, interference involving the same hand as the main task triggered WM updating, evidenced by a stronger attraction bias of the target color toward the interfering color (H3). Further exploratory analyses showed increased frontal theta activity in the matched hand condition. This activity was negatively correlated with the attraction bias on a single-trial level. We suggest that WM updating with irrelevant sensory information is triggered automatically given the action-based relevance of interference; however, a control mechanism, reflected in frontal theta, actively responds to mitigate the impact of interference.