Researchers combined data from 66 studies to clarify how the structure of the brain's white matter differs in people with autism spectrum disorder (ASD), a developmental condition affecting communication and behavior. The work was published in the journal Biological Psychiatry: Cognitive Neuroscience and Neuroimaging.
White matter consists of bundles of nerve fibers connecting different areas of the brain. It is studied using diffusion tensor imaging (DTI), an MRI method that tracks the movement of water molecules in tissue and uses it to infer fiber organization.
The authors, led by Christie D. Yoon of the University of Wisconsin-Madison, searched for papers in four scientific literature databases. The final analysis included 66 studies with 2,231 people with autism and 1,856 people with typical development. The mean age of participants across studies ranged from 2 to 50 years. In total, 881 effect-size measures were considered, a statistical quantity showing how strongly the groups differ.
Two measures were compared. Fractional anisotropy reflects how consistently water moves in one direction: the higher the value, the more ordered the fibers. Mean diffusivity indicates how freely water moves through tissue overall, which may point to the density of the structure.
On average, people with autism had moderately lower fractional anisotropy in nine tracts, including the corpus callosum, which connects the hemispheres, the corticospinal tract, the thalamic radiation, the arcuate fasciculus, the inferior fronto-occipital fasciculus, the inferior and superior longitudinal fasciculi, the uncinate fasciculus and the cingulum. For the superior longitudinal fasciculus and the cingulum, the difference reached significance only after removing several atypically influential results, so the authors call for caution.
Mean diffusivity was higher in people with autism in eight tracts: the corpus callosum, the corona radiata, the corticospinal tract, the arcuate fasciculus, the inferior fronto-occipital fasciculus, the inferior and superior longitudinal fasciculi and the uncinate fasciculus. This may indicate lower tissue density, although other explanations are possible. The largest difference was found in the arcuate fasciculus, which is linked to language, but such a strong effect could have been inflated by factors unrelated to fiber structure, such as enlarged fluid-filled spaces in the brain.
Many of the differences involved long tracts connecting distant areas of the brain. This is consistent with the hypothesis that connectivity between remote regions is affected in autism. No differences were found in the cerebellar peduncles and the internal capsule, but the authors do not rule out that the method simply performs poorly in these areas.
The size of the differences depended on study characteristics. In studies with a higher mean IQ, differences in fractional anisotropy were smaller in seven tracts, and in older samples, in two. The higher the proportion of men, the smaller the differences were in several tracts. Year of publication did not affect the result.
"Our results highlight the complex, multidimensional nature of white matter microstructure alterations in ASD," the authors concluded.
Limitations: the studies examined included a disproportionately large number of men, which limits generalization of the findings to women. Participant ages did not include older people, in whom white matter naturally changes. The sample mainly represented people with higher IQ and probably lower support needs. The DTI method is nonspecific and cannot identify the biological cause of the differences. The researchers could not rule out that some effects are explained by differences in head motion during scanning. The meta-analysis was not preregistered, and the included studies were cross-sectional, so it is impossible to judge how the differences change with age.