Neuromyelitis optica (NMO) is an inflammatory disease mainly affecting optic and spinal cords, and was originally described by Devic in 1894. There has been long controversy about whether NMO is a subtype of multiple sclerosis (MS) or a distinct disease. In Japan and other Asian countries, relapsing NMO has been called as optic-spinal form of MS (OSMS), but we reported in 2002 that OSMS was heterogeneous and it comprised both typical NMO and MS with optic spinal presentation. Recently, a highly specific serum autoantibody marker, NMO-IgG, was found in the sera of Caucasian NMO and Japanese OSMS cases, and the target antigen was identified as the water channel protein aquaporin (AQP) 4. So in NMO and OSMS, similar autoimmune backgrounds were revealed. In our anti-AQP4 antibody assay using HEK293 cells transfected with human AQP4, we found that the sensitivity and specificity of anti-AQP4 antibody was 91% and 100%, respectively, which was superior to the original immunohistological assay using mouse brain slices (NMO-IgG). The titre of AQP4 antibody correlated with the length of spinal cord lesions and relapse rate. We also studied the expression of AQP4 in autopsied cases of NMO and MS and revealed that AQP4 and GFAP, an astrocytic marker protein, were completely lost at the acute inflammatory lesions surrounding immunoglobulin and complement-deposited dilated vessels, but the myelin basic protein was relatively preserved. Those results suggest that astrocytic damage associated with autoimmunity to AQP4 may be involved in the pathogenesis of NMO, which is distinct from MS, primarily demyelinating disease. After the long history of confusion, NMO became clearly discriminated disease from MS. In this review, we focus on the historical changes of the disease concept and new knowledge gained from the clinical or immunological analyses of NMO.