Structural Connectivity of the Accumbofrontal Tract in Youth with and without Autism: Associations with Behavioral Phenotypes

Background:   Human neuroimaging work suggests that reward processing is affected in individuals with autism spectrum disorder (ASD); compared to neurotypical controls, individuals with ASD show altered brain activity to rewarding stimuli in nodes of reward circuitry including the ventral striatum, anterior cingulate, and prefrontal cortex (e.g., Scott-Van Zeeland 2010; Dichter 2012; Kohls 2013). The inferior portion of the VS encompasses the nucleus accumbens (NAcc), a region that animal studies indicate is critical for the formation of positive associations with social rewards (Dolen 2013) and attachment behavior (Keebaugh 2015). Importantly, the presence of an accumbofrontal white matter tract has recently been reported using diffusion tensor imaging (DTI) in humans (Karlsgodt 2015). This tract connects the NAcc and orbitofrontal cortex, two areas that display altered activity during reward processing in ASD. Here, we used DTI to detect and compare accumbofrontal fractional anisotropy (FA; a measure of white matter integrity) in youth with and without ASD, and relate FA to behavioral measures associated with the ASD phenotype.

Objectives:   1) Compare accumbofrontal structural connectivity in youth with and without ASD. 2) Investigate how accumbofrontal connectivity relates to measures of ASD symptomatology.

Methods:   DTI data was acquired in high functioning youth with ASD (N=44) and typically developing youth (N=44) ages 8-17. Data analysis consisted of quality control and removal of volumes with artifacts using DTIPrep (Oguz 2015), eddy current correction, brain extraction, calculation of diffusion parameters (BEDPOSTX), registration to standard space, and probabilistic tractography (PROBTRACKX; Behrens 2007). Binary masks of the left and right NAcc (defined using the Havard-Oxford Atlas) were used as seed regions for probabilistic tractography. For each subject, 5,000 streamlines were initiated from the seed voxels using the contralateral hemisphere as an exclusionary mask; a minimum threshold of 50 streamlines was applied for inclusion in subsequent analyses. The resulting map was binarized and used as a mask to extract FA values for the left and right accumbofrontal tract for each subject, which were compared between diagnostic groups and correlated with behavioral measures of intelligence and social functioning.

Results:   In both TD and ASD participants, the left and right NAcc showed structural connectivity with the ipsilateral orbital frontal cortex. There were no significant differences in FA for either the left or right accumbofrontal track when comparing TD and ASD participants (p>0.05). In ASD youth, left (r=0.40, p=0.007) and right (r=0.50, p=0.0005) accumbofrontal FA was associated with higher scores of verbal intelligence. Higher right accumbofrontal connectivity was also associated with less severe scores on the Autism Diagnostic Observation Schedule (ADOS; Lord 2000) social (r=-0.36, p=0.02) and social/communication (r=-0.32, p=0.03) subscales in youth with ASD. There were no significant behavioral correlations in TD youth.

Conclusions:   These findings indicate that structural connectivity between the NAcc and orbital frontal cortex is intact in youth with ASD, and that higher levels of FA is associated with less severe verbal and social/communication impairment. This work suggests that structural connectivity of the reward network may be an important biomarker, helping to explain variability of symptomatology in youth with ASD.