Reduced Amygdala Volumes in Adolescents with Autism Spectrum Disorders Parallels Negative Autistic Trait-Amygdala Correlations in Typically Developing Adolescent Males

Background: Previous neuroimaging studies implicate the amygdala as key in the aberrant neural circuitry underlying the social-emotional difficulties experienced by individuals with autism spectrum disorders (ASD).  Abnormal amygdala volumes in ASD have been consistently shown in prior investigations; however, whether there is an amygdala enlargement or reduction is dependent on the age of participants.  In ASD, generally greater amygdala volumes are reported among young children while adolescents and young adults may exhibit reduced amygdala size.  Nevertheless, this amygdala volume reduction during adolescence and young adulthood remains an open question.  Furthermore, it is unclear whether there are subclinical links between normal variation in amygdala volumes and autistic traits among typically developing individuals.  Such a relationship would link the full variability in ASD behavior with the amygdala, as has been done with cortical structures, most prominently superior temporal sulcus (Wallace et al., 2012).

Objectives: The purpose of the current study is to compare amygdala volumes in adolescent and young adult males with ASD versus matched typically developing (TD) males and to assess the link between amygdala volumes and self-rated autistic traits in the TD group.

Methods: 41 males with ASD (diagnosed using DSM-IV criteria and the ADI/ADOS) and 40 TD males provided high-resolution 3 Tesla anatomic magnetic resonance imaging scans.  Groups were matched on age (ASD mean=16.75 +/- 2.84; TD mean=17.04 +/- 2.73) and IQ (ASD mean=113.27 +/- 15.09; TD mean=114.03 +/- 10.74).  The FreeSurfer image analysis suite (version 5.1) was used to derive lateralized volumes for the amygdala.  Self-ratings of autistic traits were obtained within the TD group using the Autism Spectrum Quotient (AQ).

Results: The amygdala was found to be significantly smaller in adolescent males with ASD as compared to TD males (F=5.18, p<.05), which became a trend effect after controlling for intracranial volume, age, and IQ (F=3.16, p=.08).  This group difference was particularly strong for the right amygdala (F=6.12, p<.05), even after covarying the effects of these nuisance variables (F=4.11, p<.05).  Paralleling this group difference; as self-ratings of autistic traits increased, amygdala size decreased (r=-.55, p<.01) in the TD group, and this association held after partialling intracranial volume, age, and IQ (pr=-.41, p<.05).  This correlation was found on both the left (r=-.55, p<.01) and right (r=-.46, p<.01) sides, though after covariation of the nuisance variables, only the left side remained significant (pr=.42, p<.05), while the right side dropped to trend levels (pr=-.31, p=.08).

Conclusions: The present study showed that amygdala volumes are indeed reduced among adolescents and young adults with ASD (in contrast to the reported enlargement during early childhood).  Extending the literature, we found that as self-ratings of autistic traits increased, amygdala volumes decreased in the TD group.  This pattern is similar to prior cortical (thickness) findings of other key ‘social brain’ structures (e.g., superior temporal sulcus) in TD populations (Wallace et al., 2012).  By showing both ASD group reductions in amygdala volumes and negative correlations between subclinical ASD behavior and amygdala size in the TD group, these findings further implicate the amygdala in the pathophysiology of ASD.