Author response: The autism "epidemic": Ethical, legal, and social issues in a developmental spectrum disorder.

We thank Machado et al. for the comments on our review,1 and for the observations of abnormal anatomic and functional connectivity using diffusion-weighted MRI (DW-MRI). Innumerable studies have compared patients with autism spectrum disorders (ASD) and nonaffected controls at rest or while performing particular tasks, most often in small selected subject samples. For example, studies linked autism with the retino-collicular magnocellular visual pathways that project via the pulvinar to the amygdala, occipital V1, and multiple other areas, including the dorsal occipito-parietal/midtemporal streams, which participate with the frontal eye fields in processing spatial attention, moving targets, and dynamic facial expression (with inputs from the amygdala and other limbic pathways).2 However, no results consistently demonstrated specific neural networks that are uniquely pathognomonic of autism in the brain despite the fact that broad networks were linked with deficient social cognition in autism.3 Complex behaviors, like ASD, do not have a single endophenotype linked to one discrete neocortical location.4 The critical need for more research in autism should be balanced by caution about generalizing interpretations of neuroimaging findings that suggest the dysregulation of specific neural pathways. At present, DW-MRI and similar techniques, such as diffusion compartment imaging, are considered beneficial for research but not for routine clinical diagnostics.