receptor (n=5), myelin oligodendrocyte glycoprotein (MOG, n=5), aquaporin-4 (AQP4, n=1), and voltage-gated potassium channels (n=3). Because the etiology of brainstem encephalitis is widely unknown, the results described by Hacohen et al. contribute important knowledge to immunological aspects of the disease in pediatric patients.Although the numbers of patients with antibodies were small in each group, they presented with typical clinical phenotypes. NMDAR antibodies were only found in the encephalitis group, and the association of these antibodies with psychiatric features, movement disorders, or seizures -which are the main symptoms characteristic for autoimmune encephalitis -fits with the typical presentation of NMDAR encephalitis.3 In contrast, MOG antibodies were seen exclusively in patients with white matter involvement such as clinical isolated syndrome or acute disseminated encephalomyelitis (ADEM). The low percentage of MOG antibodies in this ADEM group compared to a recent study 4 prospectively investigating clinical and neuroradiological courses of ADEM patients, which identified MOG antibodies in 58% of the patients, might be explained by the overrepresentation of BBE patients.Compared to previous studies, Hacohen et al. observed different frequencies of antibody distributions, which could be explained by the characteristics and limitations of their current study. The cohort they investigated is a clinically heterogeneous population of pediatric brain stem encephalitis patients, which corresponds well to a 'real life setting' in a referral center. Nevertheless, the high frequency of GQ1b antibodies among patients with confirmed Bickerstaff's brainstem encephalitis (43%) has to be, at least in part, attributed to a selection bias, because serum samples investigated here were initially collected for GQ1b antibody testing. As pointed out by the authors, further limitations such as testing a purely Japanese cohort and the lack of cerebrospinal fluid might yield skewed results.Group sizes of each antibody found by Hacohen et al. were small, so statistical analysis has limited value for any implications on treatment strategies. However, two patients with NMDAR antibodies and persisting severe learning difficulties suggest the need for a more aggressive immunotherapy in those patients. The patient with AQP4 antibodies described here parallels a recently identified case of brainstem syndrome presenting with AQP4 antibodies, 5 which suggests the need for laboratory testing for AQP4 antibodies in patients with brainstem involvement.In sum, for pediatric patients with CNS inflammation involving the brainstem, one should be aware of possible contributions of antibodies to glial and neuronal antigens. By reasonably interpreting characteristic clinical symptoms, the identification of relevant antibodies at onset (in particular NMDAR, AQP4, and MOG antibodies) could influence treatment strategies towards a more or less intensive immunotherapy.