Purpose <p>Neuropsychiatric systemic lupus erythematosus (NPSLE) is the second leading cause of morbidity and mortality in systemic lupus erythematosus (SLE). The pathogenesis of NPSLE is poorly understood and diagnosis is difficult. Advances in neuroimaging show promise, but no patterns specific for the disease have emerged. The identification of such patterns may be facilitated by statistical techniques such as cluster analysis.</p> Materials and methods <p>FDG PET-CT scans of 46 patients confirmed with NPSLE were included. Median values of each brain region of the Brainnetome atlas were extracted after normalisation for global counts. Hierarchical clustering of patients was performed and demographic, clinical and laboratory characteristics of the clusters were compared.</p> Results <p>Our analysis revealed an optimal cluster number of two. Comparing imaging features of scans in patient cluster 1 (<i>n</i> = 23) and patient cluster 2 (<i>n</i> = 23), scans in the latter group demonstrated significantly higher metabolism in basal ganglia, thalami, medial temporal lobes and cerebellum and hypometabolism in the frontal, parietal, occipital and lateral temporal lobes. Patients in cluster 2 had higher SLEDAI-2&#xa0;K disease activity scores and erythrocyte sedimentation rates, greater frequencies of psychosis, cerebrovascular accidents, and vasculitis as manifested by rash and mucosal ulcers, and higher rates of positive anti-dsDNA and anti-Smith antibodies, as well as higher anti-Smith antibody titres.</p> Conclusion <p>We identified a subgroup of NPSLE patients with a distinct FDG uptake pattern, higher disease activity and higher incidence of certain clinical symptoms and laboratory findings. This metabolic pattern is similar to that described in autoimmune encephalitis and may be driven by a common pathological mechanism.</p>