32386
Subgrouping of Resting-State Limbic System Connectivity in Co-Occurring Autism and Social Anxiety Disorder
Anxiety disorders occur in 42-85% of individuals with an ASD (Joshi et al., 2013). The limbic system (LS) is associated with the regulation of fear (Anad et. al 2005) and previous work has found that LS connectivity at rest correlates with anxiety (Mucci et. al, 2018). Social anxiety disorder (SAD), which is characterized by social fears, has been associated with hyperconnectivity of LS-related regions (Arnold et. al, 2014). Considering the high comorbidity of SAD in ASD, understanding heterogeneity in LS connectivity may elucidate pathways for tailored treatment and response. However, no studies have examined LS connectivity transdiagnostically, including individuals with comorbid Autism Spectrum Disorder (ASD+SAD), Moreover, it is not known whether individuals with ASD+SAD show similar patterns of connectivity in the LS compared to those without ASD.
Objectives:
To elucidate heterogeneous patterns of LS connectivity across adolescents with varying degrees of social impairment. Specifically, we used a graph theory approach, Group Iterative Multiple Model Estimation (GIMME) with community detection on fMRI resting state data to examine whether LS connectivity can robustly differentiate SAD, ASD+SAD, and controls (CON). Secondary analyses examine the relationship between the LS and measures of autism severity and social anxiety symptoms. We predicted that LS subgroups will emerge such that scores on measures of ASD and SAD severity will be distinct among diagnostic classifications.
Methods:
Adolescents with ASD+SAD, SAD only, and CON participated in this study (20 per group; N=60). Participants were matched on age (M=15.19) and FSIQ (M=108.38). Diagnoses were confirmed with semi-structured clinical interviews. fMRI data were collected on a 3T TIMTrio. Functional regions of the LS were identified via Neurosynth (Yarkoni, 2011) and included: bilateral amygdala, thalamus, anterior cingulate cortex, and bilateral orbitalfrontal cortex. GIMME was used with resting-state data from each subject. Subsequently, a community detection algorithm was used to identify subgroups characterized by LS connectivity patterns (Gates et al., 2014). Autism severity was measured via self-report on the Social Responsiveness Scale (SRS-2). Social anxiety was measured via self-report with the Leibowitz Social Anxiety Scale-Child Adolescent (LSAS-CA).
Results:
Analyses revealed two distinct subgroups based on LS functioning (Fig. 1). Group one represents a hyperconnected group, whereas Group 2 displayed hypoconnectivity among the LS. Additional analyses revealed no relationship between autism severity and social anxiety symptoms between the two groups (ps>.05; Table 1).
Conclusions:
The data did not support the hypothesis that self-reported ASD or SAD symptoms are related to distinct neural signatures. Findings indicate the development of the LS may not be related specifically to clinical symptoms or diagnostic classification. Rather, results indicate that although symptom expression may be consistent, underlying neural connectivity in the LS connectivity may better explain heterogeneity. This finding holds importance for the development individualized, and neuroscience-informed interventions.