Based on epidemiological evidence that viral infection of pregnant women can strongly increase the risk for autism in their offspring, a mouse model of maternal immune activation has been established. Pregnant mice are given a respiratory infection at midgestation, and the offspring display a series of behavioral abnormalities consistent with those seen in autism, including deficits in social interaction and investigation of novel objects, as well as increased anxiety under mildly stressful conditions. Moreover, these offspring exhibit a specific type of neuropathology that is often seen in autism: spatially localized deficit of Purkinje cells in the cerebellum. Now that the mouse model has shown behavioral and histological abnormalities consistent with those found in autism, it is important to extend this work to primates. These animals possess a behavioral repertoire, such as the expression and interpretation of facial expressions, which is very homologous with human social behavior. Non-human primates therefore provide a sensitive indicator of impairments in social behavior and communication. A valid primate model of autism would also provide a unique resource for evaluating therapeutic interventions.