This is an exciting time for neuroimmunology. Over the last 10 years, several clinical syndromes have been shown to associate closely with highly specific autoantibodies directed at neuronal or glial antigens, including voltage-gated potassium channel complexes [now known to contain two proteins, Lgi1 and Caspr2 that are the main targets (Irani et al. , 2010; Lai et al. , 2010)], aquaporin-4 (Lennon et al. , 2005) and a miscellany of ionotrophic and metabotropic receptors, most importantly N -methyl-d-aspartate (NMDA) receptors (Dalmau et al. , 2008). The clinical syndromes may involve the limbic system, basal ganglia, cerebellum, brainstem, optic nerves or spinal cord (Graus et al. , 2010). Unlike the well-established antibody-mediated diseases of the peripheral nervous system or neuromuscular junction, myasthenia gravis and Lambert–Eaton myasthenic syndrome, in which the pathogenic roles of the antibodies were demonstrated relatively easily by passive transfer (Toyka et al. , 1975; Lang et al. , 1983), similar experiments with CNS disorders are technically more demanding. Some headway has been made in neuromyelitis optica by demonstrating that systemic delivery of purified aquaporin-4 antibody-positive immunoglobulin, that binds predominantly to astrocytes, results in neuromyelitis optica-like inflammatory lesions in the brain or spinal cord of mice, if the blood–brain barrier is compromised by encephalotogenic T cells (Bennett et al. , 2009; Bradl et al. , 2009; …