Exploring the Neural Correlates of Repetitive Behavior in Babies with Autism

Background: Restricted and repetitive behaviors (RRBs) are central to the behavioral phenotype of autism spectrum disorder (ASD). Once believed to emerge after core social symptoms, recent evidence suggests that atypical RRB may be an early emerging symptom of ASD (Elison et al., in press; Wolff et al., 2014). Separate work involving both human and non-human animal models indicates that repetitive behaviors, from primarily motor forms to more complex patterns of behavior, are differentially linked to parallel but functionally distinct cortico-striato-thalamo-cortical circuitry (for a review, see Langen et al., 2011). It is yet unknown whether or how this putative neural circuitry may underlay early emerging repetitive behaviors among infants who develop ASD.

Objectives: To test the predictive relationships of white matter microstructure in select fiber pathways at age 1 year, measured using diffusion tensor imaging, and repetitive behavior at age 2 years in a prospective, longitudinal sample of children with ASD.

Methods: The present study included 32 infants meeting clinical best-estimate criteria for ASD at age 2 years.  Diffusion tensor brain imaging data were collected during natural sleep as part of an ongoing study. White matter pathways of interest were deterministically segmented and microstructure characterized by fractional anisotropy (FA), a measure reflecting magnitude of directional diffusion based on tensor shape (Verde et al., 2014). Pathways examined included thalamo-frontal, cortico-striatal-spinal, and midcerebellar pathways. A mean value was created for the former two bilateral pathways. Repetitive behaviors were characterized at 24 months of age using the Repetitive Behavior Scales, Revised (RBS-R; Bodfish et al., 2000). Predictive relationships between targeted white matter pathways and repetitive behaviors were examined using a hierarchical regression model controlling for age, sex, and scan site.

Results: Results from the primary analysis are presented in Table 1. Control variables alone did not significantly predict repetitive behavior inventory at age 2. There were no significant additive effects for either the thalamo-frontal or cortico-striatal pathways. The addition of the midcerebellar pathway resulted in a significant model change, p = .002. To further explore this result, we conducted two follow-up analyses. First, we calculated two gross divisions of repetitive behavior (lower and higher-order) based on the RBS-R. The midcerebellar pathway was significantly correlated with lower (r = .49, p = .005) but not higher order (r = .18, p = .36) repetitive behavior. We then examined an additional cerebellar pathway, mean bilateral superior cerebellar peduncles. This too was uniquely correlated with lower (r = .41, p = .02) but not higher-order (r = .26, p = .17) RRB.

Conclusions: These findings suggest that structural connectivity of the cerebellum in infancy may underlie later repetitive behavior in children with ASD. This conclusion is supported by converging data implicating both the midcerebellar pathway (linking pontine nuclei and supporting cortico-cerebellar feedback) and superior cerebellar peduncles (the primary afferents to midbrain/thalamus). It is possible that repetitive behavior in infancy is initially associated with primary sensorimotor circuitry, with cotrico-striato-thalmo-cortical circuits assuming (dys)regulation of RRB through an experience-dependent process of cortical-subcortical development.