An increasing number of epidemiologic studies have implicated in utero exposure to infection in the etiopathogenesis of schizophrenia. Recent work has capitalized on the use of prospectively acquired data on infection based on maternal biomarkers. These studies suggest that several maternal infections, including rubella, influenza, toxoplasmosis, herpes simplex virus/other genital-reproductive infections, and elevations in the cytokines interleukin-8 and TNF-α, are associated with increased schizophrenia risk among offspring. Animal models of in utero infection offer the potential to corroborate these findings under controlled conditions and address etiopathogenic mechanisms. Models of maternal immune activation (MIA) and behavioral and brain anomalies in schizophrenia have primarily employed three agents: polyinosinic:polycytidylic acid (poly I:C), lipopolysaccharide, each of which mimic viral and bacterial infections, respectively, and direct maternal inoculation with influenza. Advantages and disadvantages of these approaches are discussed and the findings are reviewed. Each of these MIA models has yielded evidence, reviewed in detail, for neurochemical, neuropathologic, and behavioral anomalies in offspring that are consistent with phenotypes found in patients with schizophrenia. In addition, putative mechanisms by which MIA causes these phenotypes, the role of gestational timing of the insult, and long-term modulation of the immune system following prenatal immune challenge are discussed. Finally, the “back translation” of these findings to the design and implementation of new epidemiologic studies aimed at further refining and testing this hypothesis is discussed.