Of Many Autisms or One Neurodevelopmental Disorder

Defining diagnostic boundaries is difficult. There is a case to be made for lumping disparate neurodevelopmental disorders, including autism, ADHD, and OCD, into a single category. Conversely, there are strong arguments for splitting autism into multiple, narrower diagnoses focused on genetics or other aetiological factors. Here I will use brain imaging data from both large human population samples as well as from over 100 mouse models related to the disorder to explore both the case for lumping and splitting. Human imaging finding, coming from structural MRI, diffusion imaging, and functional MRI indicates that the boundaries between different neurodevelopmental disorders are hard to recover in brain signatures. The mouse data indicates that there are 4 separate classes of models which share a neuroanatomical phenotype and link with known genetic pathways. The human and mouse data together suggest that our current nosological categories do not have a strong basis in biology, and that future choices of where to lump and where to split will be context dependent.