Abstract
BACKGROUND: Despite high rates of autism spectrum disorder (ASD), understanding of pathophysiology is limited. The RAS-mitogen-activated protein kinase (RAS-MAPK) pathway plays a crucial role in ASD and is altered in children with Noonan syndrome (NS). Children with NS offer a unique model to disentangle genetic and neurological underpinnings of ASD. METHODS: This study aimed to examine functional brain network anatomy underlying ASD symptoms in children with NS (n=28, mean age=8.24), and tested generalizability of models developed in a non-syndromic cohort enriched for ASD (Autism Brain Imaging Data Exchange (ABIDE), n=352, mean age=11.0). Connectome-based predictive modeling (CPM) was applied to fMRI data to predict the severity of autism symptoms, indexed by the Social Responsive Scale (SRS), in children with NS. Next, we tested if a model developed to predict autism symptoms in an autism-enriched sample of children without genetic diagnosis (ABIDE) could predict autism symptoms in children with NS. RESULTS: Predicted SRS scores were significantly associated with observed SRS scores in NS (r (s) =0.43, p=.011). Application of the predictive model generated in the autism-enriched cohort (ABIDE) significantly predicted observed SRS scores in NS (r (s) =0.460, p=.018). Predictive brain networks in both NS and the non-syndromic cohorts included subcortical-cerebellar networks and visual processing networks. LIMITATIONS: The size of our NS cohort is small, given the rarity of NS. However, the significant cross-dataset comparison yielded in this study suggests that use of large publicly available datasets can be useful in contextualizing smaller and harder to collect datasets in rare genetic syndromes. CONCLUSIONS: The presence of shared brain networks suggests a converging pattern of functional connectivity underlying autism symptoms, irrespective of genetic diagnosis. Evidence of shared brain networks in children with idiopathic autism and NS highlights the role of RAS-MAPK in autism symptoms and points to the value of leveraging human genetic models to enhance our understanding of idiopathic ASD.