Ali Yılmaztekin¹ (ali.yilmaztkn@gmail.com), Eren Günseli¹; ¹Sabancı University

Working memory (WM) is vulnerable to interference, whereas repeated exposure often reduces this vulnerability by strengthening long-term memory (LTM) traces. Prior work has shown that context changes can trigger active representation of LTM items in WM. However, it remains unclear whether such reactivation reinstates fragile WM representations or robust LTM-like states. Across two experiments, participants repeatedly studied and reported the color of one (Exp 1) or two (Exp 2) memory items over six consecutive trials. On half of the 5th color repetitions, the background context was changed at encoding to trigger reactivation in WM (Özdemir et al., 2024). A subset of these trials also featured a delay-period interference task. Participants either located an odd-colored item from a circle of images (Exp 1) or checked for a repeated color in a grid of colors (Exp 2). In both experiments, memory errors decreased across color repetitions. The interference task increased memory errors for the first color repetition, but this interference cost diminished with subsequent repetitions. Importantly, a context change on the 5th color repetition did not increase the interference cost, which would be expected if reactivation reinstated vulnerable WM representations. These findings suggest that LTM traces continue to buffer against disruption even when they are reactivated in WM. Alternatively, reactivation may reinstate a representational format that is less vulnerable to interference than newly encoded information.