Gene expression data are most useful if they can be associated with specific cell types. This is particularly so in an organ such as the brain, where many different cell types lie in close proximity to each other. We used zebra finches (Taeniopygia guttata), fluorescent tracers and laser capture microdissection (LCM) to collect projection neurons and their RNAs from two interspersed populations from the same animal. RNA amplified from each cell class was reverse transcribed, fluorescently labeled, and hybridized to cDNA microarrays of genes expressed in the zebra finch brain. We applied strict fold-expression criteria, supplemented by statistical analysis, to single out genes that showed the most extreme and consistent differential expression between the two cell classes. Confirmation of the true expression pattern of these genes was made by in situ hybridization and Taqman quantitative PCR (qPCR). High quality RNA was obtained, too, from backfilled neurons birth-dated with bromodeoxyuridine (BrdU). We also quantified changes in the levels of three genes after singing behavior using qPCR. Thus, we have brought together a combination of techniques allowing for the molecular profiling of intermingled populations of projection neurons of known connectivity, age and experience, which should constitute a powerful tool for CNS research.