One big problem in conducting studies on the causes of autism is reliable diagnosis. There are no biological markers of autism so autism is inferred from the child's behavior. The problem arises then that there such obvious variably from one child with autism to another that a "spectrum" is required to "explain" this variability. So is autism one disorder or many? Obviously, this is a fundamental question that must be answered before much progress can be made in discovering the cause of autism and developing more effective therapies. A study at the University of California at San Diego found differences in functional magnetic resonance image (fMRI) scans of children diagnosed behaviorally with autism and normal controls. The fMRI's, taken while the children were sleeping, found that the two hemispheres of the brains of the children with autism were not "syncing." Well, if this is replicated in other studies it will certainly advance our understanding. The only downside I see is that research will then be done in large universities with large grants. The cost for a fMRI and a MRI suite can run $500,000.