With the rise of e-Learning in engineering education, understanding the impact of individual differences on the ways students communicate and collaborate on-line has become increasingly important. The research described here investigates the influence of cognitive style on the interactions within student social networks in an on-line learning environment, with a particular focus on student engagement, patterns of communication, and the self-directed creation of sub-groups (i.e., cliques). The Kirton Adaption-Innovation Inventory (KAI) was used to assess cognitive style, and UCINET software was used to analyze the interactions of two cohorts of Systems Engineering students throughout a series of asynchronous on-line discussion forums across two graduate-level courses. Among the findings, the highly heterogeneous style composition of the cliques formed by the students suggests that e-Learning environments may mask cognitive differences that have been shown to create conflict in face-to-face student interactions. Links between cognitive style, expansiveness, influence, leadership, and students' choices between resident and on-line programs are also discussed.