Co-Occurring Anxiety Disorders Are Uniquely Associated with Decreased Amygdala Volume in ASD

Background: Although deficits in amygdala function have long been associated with autism spectrum disorder (ASD), findings of abnormal amygdala structure have been less consistent.  The most recent data indicate that amygdala undergoes early overgrowth in ASD followed by attenuated growth (i.e., comparable to controls) beginning in early adolescence. However, research to date has largely failed to consider the influence of co-occurring anxiety disorders, which affect more than 40% of the ASD population and are associated with abnormal amygdala volume (including findings of both increased and decreased size). Individuals with ASD and co-occurring anxiety may have a unique trajectory of amygdala development that has been consistently overlooked.

Objectives: This study tests the hypothesis that individuals with ASD and co-occurring anxiety show abnormal amygdala structure (as indexed by amygdala volume).

Methods: 53 participants with ASD and 37 typically developing controls (TDCs; mean age = 12 years for both groups) underwent extensive diagnostic assessment and MRI scanning.  ASD was assessed using the Autism Diagnostic Observation Schedule and either the Autism Diagnostic Interview-Revised or DSM-driven parental interview.  The presence or absence of an anxiety disorder was established using the Anxiety Disorder Interview Schedule.  All participants completed a high-resolution T1-weighted structural MRI scan.  ASD participants were categorized based on the presence or absence of a DSM anxiety disorder.  The resulting three final groups (ASD+Anxiety [n = 28], ASD-Anxiety [n =25], and TDC [n = 37]) were matched on age and intellectual ability (measured via the Differential Abilities Scale - II).  The ASD+Anxiety and ASD-Anxiety groups were matched on core ASD symptoms (measured via the Social Communication Questionnaire). Amygdala volumes were estimated using the FSL program FIRST, then manually inspected (data for two participants were discarded at this stage due to poor segmentations). Group differences in amygdala volume were estimated via a general linear model including total brain volume as a covariate.

Results: When covarying brain volume, the ASD+Anxiety group showed decreased right amygdala volume compared to the TDC and ASD-Anxiety groups (p-values of .035 and .003, respectively).  The ASD-Anxiety and TDC groups did not differ in right amygdala size, nor were any group differences found in left amygdala volume.  Although not predicted a priori, the ASD+Anxiety group also showed increased brain volume relative to the TDC group (p = .056), but not to the ASD-Anxiety group.

Conclusions: Individuals with ASD and co-occurring anxiety show a distinct amygdala abnormality that is not shared by individuals with ASD but non-clinical levels of anxiety.  The present finding of decreased amygdala volume in ASD and anxiety is novel and consistent with some previous studies of anxiety disorders.  On the other hand, findings of comparable amygdala volume between the ASD group without anxiety and the TDC group are consistent with previous studies showing null effects of amygdala volume in early adolescence. Brain volume results indicate that this relative decrease may need to be interpreted in the context of broader patterns of abnormal brain development that are unique to individuals with ASD and anxiety.