Human autoantibodies against the NMDA receptor, aquaporin-4 and CASPR2 cause a set of clinically-distinctive central nervous system illnesses. In this seminar, I aim to provide an overview of my lab’s work in aiming to deconstruct the clinically-relevant B cell immunology of these conditions. First, I will discuss the potential contribution of germinal centres versus long-lived plasma cells in the generation of the autoantibodies, and show data supporting the patient tumours as ectopic germinal centres. Next, I will highlight our efforts to better characterise the relative roles of pre- versus post-germinal centre B cells in the generation of the autoantibodies, and, finally, describe our single cell techqniues which isolate cognate-paired, antigen-specific recombinant antibodies from these patients to better inform the sites of tolerance breakdown, and the relative pathogenicity of the patient antibodies through the B cell lineage.