Autism is now widely viewed as a neurodevelopmental disorder, although the underlying biological causes remain to be established. In this review, we examine the literature in magnetic resonance imaging (MRI) as applied to autism, discuss the findings that have emerged, and give directions for potential future research. To date, structural MRI results are inconsistent, partly due to the heterogeneity of the disorder itself, and partly due to the different composition and the varied degree of matching of the studied groups. However, recent studies have begun to elucidate the underlying neuroanatomical abnormalities and brain-behavior relationships in autism, with the most consistent finding being increased brain volume in autism. Future large-scale longitudinal structural imaging studies, starting at very young ages, investigating homogeneous groups of patients and extensively matched control groups, and making use of (combinations of) newer and more sophisticated techniques, hold a great promise to further elucidate the enigma of autism.