Imaging Autism in Mouse and Man

Background: Autism Spectrum Disorders (ASDs) are complex and still poorly understood. They are highly heritable, yet no single gene discovered to date accounts for more than 1-2% of known cases. Currently 250+ genes have been associated with Autism in the human population, and while ASD is associated with communication and social deficits, as well as repetitive behaviours, individual clinical presentations are highly heterogeneous.

Objectives: To identify clusters within the autistic spectrum using brain imaging as a shared endophenotype in both clinical studies and mouse models of the disease.

Methods: A cohort of over 100 children and adolescents diagnosed with autism were imaged on a 3T MRI. A separate control cohort of over 300 was also acquired on the same scanner using the same imaging protocols. Information about neuroanatomy and brain connectivity was extracted using state of the art image processing algorithms.

At the same time, over 30 mouse lines carrying mutations linked to autism were imaged on a 7T MRI, totaling over 500 scans. Data were assessed for alterations in anatomy and then further clustered into distinct groupings of mouse models sharing common anatomical phenotypes. We additionally investigated which areas of the brain were most consistently affected across all mouse models. Lastly, the mouse imaging data was integrated with the gene expression maps from the Allen Brain Institute to explore potential processes driving the imaging findings.

Results: Human imaging results comparing children and adolescents with autism to typically developing controls showed patterns that were diffuse throughout the brain and furthermore significantly more variable in the autistic cohort. The mouse data revealed that the brain areas affected were specific to each individual mouse model studied, resulting in a similarly diffuse and heterogeneous pattern of brain alterations once all the models were taken in aggregate. The most consistently affected brain regions were the corpus callosum, cortex, and cerebellum. The mouse models further clustered into three sets with shared imaging phenotypes.

Conclusions: Autism presents with marked hetereogeneity in phenotype. Using imaging of neuroanatomy and brain connectivity we identified the same variability in these endophenotypes. Combining mouse and human imaging data suggests potential stratifications of autism into clusters sharing similar phenotypes.