The brain, like many organs in the body, comprises an array of different cell types that act in concert to achieve its many functions. These cell types vary dramatically in molecular composition, morphology, spatial distribution, electrical activity, signaling, and association with sensory processing and behavior. As a result, a comprehensive investigation of molecular alterations in disease or perturbation studies needs to account for cell type heterogeneity in the brain. Bulk profiling of tissue is widely used in neuroscience, and is now a relatively mature field, with a range of experimental, technical, and computational approaches that are broadly and consistently used. One major question in bulk profiling, however, is the contribution of individual cell types to the overall bulk signature. Recently, advances in scale and sensitivity of single-cell methods have provided cell type-specific signatures in health and disease, offering a solution to the problem of inferring cell type contributions to bulk signatures. This is the main goal of bulk deconvolution, which aims to deconvolve cell type-specific signatures and proportions from bulk data, either de novo or using reference profiles obtained from single-cell data. Here, we present an overview of some basic principles of deconvolution, followed by a general workflow on applying deconvolution methods to bulk RNA-seq data in order to assess compositional differences in individual cell types that may be associated with experimental variable of interests.