Microglia are tissue-resident macrophages of the CNS that orchestrate local immune responses and contribute to several neurological and psychiatric diseases. Little is known about human microglia and how they orchestrate their highly plastic, context-specific adaptive responses during pathology. Here we combined two high-dimensional technologies, single-cell RNA-sequencing and time-of-flight mass cytometry, to identify microglia states in the human brain during homeostasis and disease. This approach enabled us to identify and characterize a previously unappreciated spectrum of transcriptional states in human microglia. These transcriptional states are determined by their spatial distribution, and they further change with aging and brain tumor pathology. This description of multiple microglia phenotypes in the human CNS may open promising new avenues for subset-specific therapeutic interventions.