Early collaboration network archetypes are associated with productivity and publishing longevity in cognitive science

Academic success is typically modelled as an individual-level outcome. Systems approaches, however, suggest that outcomes emerge from interactions between individuals and structured environments, including collaboration networks. We examined whether early-career collaboration networks shaped trajectories of success in cognitive science. It remains unclear whether cognitive science authors differ systematically in early collaboration structures and whether such differences relate to long-term productivity and publishing longevity. We analysed ~ 1.5 million publications from OpenAlex in cognitive psychology (post-2000). We randomly sampled 50,000 focal authors and constructed their early-career collaboration networks. For each focal author, we quantified connectivity, brokerage, local clustering, ties to highly connected co-authors, and collaboration stability. Unsupervised clustering identified network archetypes, and these were visualized with principal components analysis. Clusters were then compared on productivity and publishing longevity. Networks of 49,487 focal authors were modelled, revealing five distinct early-window archetypes with characteristic network signatures independent of publication volume. Archetypes differed in long-term productivity and publishing longevity. Early collaboration structures are associated with divergent publishing trajectories in cognitive science. Authors embedded in stable collaborations showed longest publishing tenure, followed by authors occupying brokerage positions spanning disconnected groups. Relatedly, these two author archetypes also demonstrated the largest volume of publications. These findings support systems-level accounts of scientific success with authors embedded in social network structure.