This systematic review aims to determine whether or not structural magnetic resonance imaging (sMRI) data support the DSM-5 proposal of an autism spectrum disorder (ASD) diagnostic category, and whether or not classical DSM-IV autistic disorder (AD) and Asperger syndrome (AS) categories should be subsumed into it. The most replicated sMRI findings in patients with ASD compared with healthy controls are increased total brain volume in early childhood and decreased corpus callosum volume. Regarding the notion of a "spectrum", some studies support that AS and AD are similar but "quantitatively different" diagnostic categories, whereas others support that they are "qualitatively different" entities with specific brain structural abnormalities. It seems that there are still not enough arguments from sMRI data for or against subsuming DSM-IV categories under a single ASD category. (Contains 3 tables.)