A meta-analysis of functional imaging studies in Autism Spectrum Disorders: the role of anterior cingulate cortex

Background: Functional imaging studies of autism spectrum disorders (ASD) have reported abnormalities in regions related to mentalizing and facial processing. A parallel line of studies has focused on the brain correlates of additional cognitive and sensory impairments related to ASD. This imaging literature is characterized by generally small sample sizes and task design differences which impede true refutation or replication of results. Objectives: To conduct an objective, quantitative meta-analysis of published functional neuroimaging studies of ASD. Methods: We used the activation likelihood estimate (ALE) technique implemented in BrainMap. ALE were calculated voxel-wise by modeling each coordinate with an equal weighting using a 3-D Gaussian probability density function (FWHM=10 mm). The resulting ALE-map was entered into a permutation test to calculate voxel-wise activation likelihood. Statistical maps were corrected for multiple comparisons using false discovery rates (p< 0.05, corrected). Thirty-two functional imaging studies comparing individuals with ASD and Neurotypical Controls (NC) were divided in 18 focusing on social processes and 14 examining non-social processes. Within ASD and NC we ran meta-analyses including 1) all studies combined, 2) social studies only, and 3) non-social studies only. To compare NC and ASD, we calculated the difference between the two group ALE maps for each study-set examined. Each map was entered into permutation analyses to generate voxel-wise statistical scores. Results: Analyses including all studies combined indicated greater probabilities for activation in NC compared to ASD in ventral and dorsal anterior cingulate (ACC). Secondary analyses revealed a dissociation in the ASD with vACC hypofunction only being observed for social-task studies, and dACC hypofunction for non-social. Conclusions: Results suggest abnormalities in two ACC-regions, one commonly implicated in social cognition, and the other in executive control. These areas are key nodes of structurally and functionally distinct networks the integrity of which should be examined in individuals with ASD.