Mounting evidence suggests that autoantibodies contribute to the pathogenesis of demyelination in the PNS and CNS. Rapid reversal of electrophysiological blockade after plasmapheresis or intravenous immunoglobulin treatment for acute or chronic inflammatory demyelinating polyneuropathy is more likely to result from removal or neutralization of an antibody that impairs saltatory conduction than from remyelination. Although up to 30% of patients with acute or chronic inflammatory demyelinating polyneuropathy harbour autoantibodies, specific antigens have been identified in no more than 13% of cases. To date, autoantigens identified at the node of Ranvier include neurofascin 186, gliomedin and possibly moesin in the nodal domain, and contactin-1, Caspr1 and neurofascin 155 in the paranodal domain. In some patients with multiple sclerosis, paranodal CNPase and juxtaparanodal contactin-2 trigger a humoral response. This Review explores the molecular anatomy of the node of Ranvier, focusing on proteins with extracellular domains that could serve as antigens. The clinical implications of node-specific antibody responses are addressed, and the best approaches to identify antibodies that target nodal proteins are highlighted. Also discussed are the roles of these antibodies as either secondary, disease-exacerbating responses, or as a primary effector mechanism that defines demyelination or axonal degeneration at the node, identifies disease subtypes or determines response to treatments.