Changes in Local Gyrification Index Across Childhood and Adolescence in Autism Spectrum Disorder (ASD)

Background:

A large body of evidence from volumetric MRI supports an atypical trajectory of early cerebral overgrowth in Autism Spectrum Disorder (ASD) followed by abnormally slow growth. Comparatively few studies have examined local patterns of gyrification, and the relationship between gyrification and measures of cortical thickness (CT) and surface area (SA) remains poorly understood. Developmentally, gyrification is thought to be influenced by mechanical forces resulting from differential rates of early cortical expansion (radial and tangential expansion, differences in laminar expansion). Atypical gyrification patterns may therefore reflect the timing of brain growth abnormalities in ASD and elucidate the history of early developmental disturbances.

Objectives:

In an earlier study, we found that the local gyrification index (lGI) decreased with age more rapidly in an ASD than a TD comparison group. The current study sought (1) to replicate these findings in an independent sample, and (2) to investigate the relationships between CT, SA, and lGI.

Methods:

T1 weighted MRI sequences (1mm³) were downloaded from the Autism Brain Imaging Data Exchange for 137 right-handed males aged 7–18 years (60 ASD, 77 TD; all from New York University). FreeSurfer (v5.3.0) was used to obtain measures of CT, SA, and lGI. Following quality assurance and group matching on age and non-verbal IQ, 31 ASD and 31 typically developing (TD) participants were included in analyses. A general linear model was used with group and age (continuous) as predictors and Social Responsiveness Scale (SRS) scores as a correlate. Correction for multiple comparisons used Monte Carlo null-z simulations.

Results:

For lGI, significant diagnosis-by-age interactions were found bilaterally in rostral middle frontal gyrus, in left inferior frontal gyrus (pars triangularis), and in right precuneus. Whereas lGI was relatively stable across age in the TD group, the ASD group exhibited a significantly greater negative slope (decreasing lGI with age). A main effect of group (ASD<TD) was also found in left medial orbitofrontal and right inferior parietal clusters. For CT, a significant group effect (ASD>TD) was found in a left lateral occipital cluster. No significant effects were observed for SA. Correlation analyses in the ASD group revealed that greater social impairment (higher total SRS) was associated with lower lGI bilaterally in fusiform clusters, after controlling for age.

Conclusions:

Declining lGI with age is often reported in typical adolescence, and here, regions of steeper age-related decrease were observed in ASD compared to TD participants in frontal and parietal lobes. A similar interaction was found in our earlier study on an independent sample, in partially overlapping cortical regions, largely within the frontal lobe. Together, these findings indicate an atypical developmental trajectory for frontal lGI in ASD. Correlations between fusiform lGI and SRS add additional support to the functional relevance of this morphometric measure. Further comparisons will be presented to directly assess associations between lGI, CT, and SA in larger samples.