Longitudinal Age-Related Changes in Cortical Thickness From Childhood to Adulthood in Autism

Background: Cortical development across the lifespan in autism is unknown.  Greater age-related decrease in cortical thickness in the occipital and temporal lobes were reported in a preliminary longitudinal investigation by Hardan and colleagues (2009) in adolescent males (aged 8-12 years at baseline) scanned twice over 30 months. We aim to expand on these findings.  

Objectives: To describe longitudinal age-related changes from childhood to adulthood in cortical thickness in individuals with autism compared to typical development.  

Methods: Freesurfer derived cortical thickness measurements (https://surfer.nmr.mgh.harvard.edu/ftp/articles/desikan06-parcellation.pdf) were examined in 98 males with autism (age range 3-36 years at first scan) and 61 typically developing males (age range 4-39 years at first scan).  Each participant was scanned 1-3 times, on average every 2.5 years, on a Siemens 3T MRI scanner (total of 353 scans).  Mixed effects models were used to describe longitudinal cortical thickness growth in the autism group in comparison to typical development. 

Results: Age-related decreases in cortical thickness were found across the entire cortex in both autism and typical development.  Significantly greater age-related cortical thinning was found in autism bilaterally in the occipital cortex (cuneus, lateral occipital, lingual, pericalcarine) and the majority of the parietal regions (left supramarginal, right precuneus, bilateral inferior and superior parietal and postcentral gyrus).  In the frontal lobe, greater age-related cortical thinning in autism was found in the left frontal pole, precentral gyrus, pars triangularis, rostral middle frontal, right paracentral, and bilateral pars opercularis.  Significant group by age interactions were found in only two temporal lobe regions. The left bank of the superior temporal sulcus and right transverse temporal gyrus showed greater thinning with age in autism. There were no group differences in longitudinal thickness changes in cingulate cortex.  The only region with significant group differences in mean cortical thickness was the left rostral middle frontal cortex, which was 2.8% thicker in autism (p=0.01; at centered age of 14.75 years). 

Conclusions: This study is the first to describe longitudinal age-related cortical thickness changes in autism from childhood into adulthood. In the absence of group mean differences for most regions (at centered age of 14.75 years), developmental differences were apparent and greater age-related cortical thinning was found in autism.  The occipital and parietal lobes were most affected, but group differences in age-related changes in subregions of the frontal and temporal lobes also emerged.  Future analyses will examine potential clinical correlates for individual differences in atypical age-related trajectories.