The cellular diversity of the mammalian brain is unparalleled by any other tissue: hundreds, if not thousands of different cell types act in concert to execute complex behavioral and physiological functions. In neuropsychiatric disease, some of these functions go awry, but the pathogenic mechanisms remain poorly understood. Characterizing how specific cell populations differ in disease-relevant contexts offers an opportunity to more clearly localize pathogenic processes, leading to better disease models and new potential treatments. The first step is to develop reliable assays that comprehensively characterize brain cell diversity. Genome-wide expression is an information-rich means of characterizing the functional states of cells. We recently developed a technology, Drop-seq, which enables tens of thousands of individual cells to be transcriptionally profiled in straightforward, inexpensive experiments. We are using this technology to build comprehensive cell-type atlases of the major tissues of the mouse brain, and to examine how cell types transcriptionally vary across both genetic and environmental contexts. These contexts include: 1) natural variation in mouse strains and (where experimentally tractable) human tissue, to understand when in development, and in what cell types in the brain, evolution is most active; 2) pharmacological manipulations connected to psychiatric disease pathogenesis and treatment. The work involves an exciting intersection of techniques in molecular and cellular neuroscience, genomics, and computational biology. In addition, we develop new technology to more deeply explore cellular variation in complex brain tissues. Currently, our lab is working on a high-throughput transcriptional profiling technique that retains spatial information about the individual cells of a tissue, and can be used on post-mortem brain bank samples. We hope this technique will further help elucidate the cellular mechanisms underlying brain diseases whose pathogeneses remain mysterious.