Structural Brain Differences between Autistic Children and their Typically-Developing Siblings: a Voxel-Based Morphometry Analysis

Background: Morphologic brain differences have been identified between children with autism and typically-developing children, though findings have been inconsistent across studies. One potential cause is the variability in brain morphology resulting from the influence of multiple genetic and environmental factors on brain development. A strategy to control for such factors is to compare brain structure between siblings discordant for autism.

Objectives: To assess the relationship between brain morphology and autism using voxel-based morphometry to examine siblings discordant for autism.

Methods: Participants included 27 same-gender sibships consisting of one child with autistic disorder (AU; confirmed with ADI-R and ADOS-G) and one typically-developing sibling (TD). Subjects were between 6-13 years old (AU 9.7±1.7; TD 9.1±1.9), and all pairs were less than 4 years apart. High-resolution structural magnetic resonance images were pre-processed (including modulation) and analyzed (paired t-test covarying for age, gender, and total gray/white matter volume) using SPM5 and VBM5.1 (p=0.01 corrected).

Results: AU had greater grey matter volume than TD in the right insula and posterior perisylvian region and the left insula and inferior temporal lobe. AU had greater white matter volume in the right temporal stem and inferior temporal lobe. AU had less grey matter volume than TD in bilateral anterior cingulate cortex, and less white matter volume in posterior corpus callosum and cingulate gyri and left anterior parietal lobe.

Conclusions: Volume differences found in temporal lobes, insulae, and cingulate cortices are consistent with differences found in some prior structural and functional studies of autism, and correspond to areas involved in language and social behaviors known to be abnormal in autism. Using autism-discordant siblings, we controlled for many environmental and (non-autism-causing) genetic factors which likely confounded many prior studies. These results are therefore likely more robust than those found previously. We recommend this strategy for use in future structural neuroimaging studies of autism.