Cerebellar Activation Differentiates Children with Autism and Siblings in a Static Face-Processing Task

Background:  

While face-processing deficits are known throughout Autism research, the exact mechanism is unknown. Prominent theories include problems in attentional modulation, global spatial processing, and errors in neural machinery. Knowledge of what the exact cause is may be essential to finding an effective intervention method.

Objectives:  

Our objective was to examine what areas of the brain showed specific activation for various categories of visual stimuli between typically developing adults, children with Autism Spectrum Disorder, and unaffected siblings. 

Methods:  

Adults (n = 23), children with Autism (n = 22), and unaffected siblings (n = 27) were presented black-and-white visual stimuli containing either faces, houses, vehicles, letters, or numbers in a passive-viewing task. Regions of interests were established from adult data using a GLM and by using a conjunction analysis (e.g. face> house^object^letter^number), then extracting beta values. Average beta values were then compared between children with Autism and siblings using pairwise t-tests. In a novel analysis method, the inner product between vectorized representations beta weights for face, house, and object stimuli were computed between children with Autism and unaffected siblings relative to a canonical ‘adult’ vector, yielding a singular measure of multivariate similarity. A one-way ANOVA was then run on these data, and correlations with respect to age and IQ were calculated.

Results:  

Of the face-selective regions found in the conjunction analysis, only one area—the left inferior semilunar lobule of the cerebellum—showed a significant difference in activation between children with Autism and unaffected siblings. The vector analysis yielded a similar result—of all the areas, only this cerebellar region showed a significant difference in similarity: siblings were significantly more similar to adults than children with Autism. The ANOVA on this region showed a group by condition interaction as well. Across all house-selective regions, no single area showed significance in the vector analysis, but when combined, children with Autism showed significantly more similarity than did siblings, suggesting a preserved house-network in Autistic children. Finally, there was a significant negative correlation between IQ and similarity to adults in unaffected siblings.

Conclusions:  

It is clear that the cerebellum requires more attention as a social mechanism. Its functional connectivity to the thalamus and striatum (which were also implicated as face-regions), as well as other work linking the cerebellum with attention, leads us to believe that the left inferior semilunar lobule may be an area for social attention. To further test this hypothesis, we are building a study examining 2-dimensional versus (simplistic) versus 3-dimensional (complex) facial stimuli, using the fusiform face area, amygdala, and inferior semilunar lobule as seed regions. Our goal in this is to further differentiate cerebellar function and to target a possible method of intervention using cartoon stimuli.