Societal structures and memory organization models share network-like features, offering insights into how information spreads and shapes collective memories. In this study, we manipulated the structure of lab-created community networks during a computer-mediated recall task using the Deese-Roediger-McDermott paradigm to test the spreading activation theory of true and false memory formation. We hypothesized that social network structure, whether clustered or not, would influence memory accuracy. Our results showed that clustered networks reinforced true memories by promoting mnemonic convergence, while non-clustered networks led to more false memories by increasing widespread cross-activation. These findings highlight how social network topology impacts memory dynamics and collective knowledge evolution.