Activation Likelihood Estimation Guided Investigation into the Anatomy of the Social Brain in Autism

Background:  Impairment in social communication is a hallmark feature of autism spectrum disorder (ASD) (DSM-V). Functional and anatomical alterations found in a set of brain areas, called the “social brain” may underlie these deficits in autism. These regions include: the temporal-parietal junction (TPJ), precuneus, superior temporal sulcus (STS), anterior cingulate cortex (ACC), fusiform gyrus (FG), and medial and inferior aspects of the pre-frontal cortex (MPFC, IFG) (Ashwin et al., 2007; Blakemore, 2008; Frith, 2001; Hadjikhani et al., 2007; Pelfrey et al., 2011). However, heterogeneity in ASD, variability among fMRI tasks, and differences in participant age have made the identification of a social brain signature of autism challenging.

Objectives:  The goal of this study is to characterize the functional differences in the social brain in individuals with ASD using activation likelihood estimation (ALE), and to use these results to examine the neuroanatomy of social brain in this population.

Methods:  Statistically significant foci of brain activity from 50 fMRI studies comparing individuals with autism to TD controls on 16 different types of social cognition paradigms (e.g., face processing, theory-of-mind, emotion processing) were entered into GingerALETM using a cluster-level threshold of p < .05. This included 675 individuals with ASD and 695 typically developing (TD) controls (mean age range 9-33).  A mask of the results was created using the ALE results, which then was mapped to the pial surface of an average brain template using structural MRI from 115 individuals (55 ASD and 60 TD) using FreesurferTM’s bbregister, and performed cluster-level analyses on group differences in brain volume, surface area, and thickness in each of these ROIs.

Results:  ALE - Statistically significant clusters in the cortex included the right fusiform/lateral occipital gyrus (k = 17480mm³), IFG (k = 16760mm³), middle and transverse temporal gyrus (k = 11736mm³), and precuneus (k = 880mm³), and left fusiform/middle occipital gyrus (k = 15608mm³), STS/MTG (k = 7552mm³; k = 1104mm³), insula (k = 5264mm³), MPFC (k = 2480mm³), IFG (k = 4408mm³), and postcentral gyrus (k = 2112mm³). Morphometry - Decreased cortical surface area was found in the left STS/STG in our ASD participants relative to intracranial surface area (ICSA) (k = 204.05mm²), and age relative to ICSA (k = 135.98mm²), with additional decrease in the right insula relative to ICSA (k = 65.7mm²).  Increases in the fusiform and the IFG in ASD were found as a function of mean cortical thickness (k = 65.51mm²) and age (k = 64.57mm²).  

Conclusions:  Our meta-analysis of social cognition studies highlights a network of brain regions that are part of the social brain.  Anatomical mapping of these regions suggest a subsection of regions to be significantly altered in ASD. Some of these regions, STS and fusiform, have been previously proposed as potential neuroendophenotypes of autism (Kaiser et al., 2010; Spencer et al., 2011). Age-related effects suggest that age and total-brain measures are important when considering cortical morphology. In addition, our findings suggest social brain alterations in ASD at functional and anatomical levels reflecting the complexity of the disorder.