Atypical Anatomy of Primary Visual Cortex and Links with Intelligence in Children with Autism Spectrum Disorders

Background: Previous research has indicated that individuals with Autism Spectrum Disorders (ASDs) show widespread abnormalities in cytoarchitecture and MRI derived measures of neuroanatomy, including cortical thickness (CT) and surface area (SA). However, primary visual cortex (V1) has received relatively little attention. Functional MRI findings, on the other hand, have shown atypical recruitment of V1 during performance on language and cognitive tasks in ASDs. In a previous study on a smaller, partially overlapping sample, our group found differences in functional connectivity of V1 in comparisons of ASD subgroups with high (>115) vs. low (<85) IQ.

Objectives: To characterize CT and SA of V1, and relate these measures to intellectual functioning in children with and without ASDs.

Methods: High quality T1 MRI scans from 165 children [83 ASDs, 82 Typically Developing (TD)], ages 7-18 [mean age(standard deviation) =13(3), mean IQ = 108(13), range: 66-141)], from two ABIDE sites (San Diego State University and New York University), were included. Groups were matched on age, sex, handedness, and IQ. After standard preprocessing and rigorous quality control of images, CT and SA of left and right pericalcarine (V1) cortex, and left and right hemisphere were calculated using Freesurfer. Linear regression was used to test group differences in CT and SA, as well as group by IQ interactions; relationships between age and CT and SA, and group by age interactions were also examined. In all analyses, we controlled for scanning-site and average whole-hemisphere CT or total SA.

Results: The ASD group had significantly decreased left V1 SA compared to the TD group (p = .002). In contrast, V1 CT was significantly increased bilaterally (left: p = .03; right p =.02) in the ASD, compared to the TD group. Diagnosis significantly moderated the relationship between IQ and V1 SA, bilaterally (left: p = .01; right p =.02). There was no such moderating effect for CT. ASD symptom severity (ADOS scores) was not significantly related to V1 SA or CT. Finally, independent of diagnosis, age significantly predicted an increase in right V1 SA (p = .004), and decrease in right V1 CT (p < .001); however, after controlling for whole-hemisphere SA/CT these effects were not significant.

Conclusions: In congruence with the literature on atypical function of visual regions in ASD, we found abnormal anatomy of V1 (reduced SA, but increased CT). Our findings add an anatomical basis to the growing evidence of a ‘special status’ of visual functions in behavioral profiles of intelligence in ASDs. In ASDs, but not in matched TD controls, SA of V1 was positively associated with IQ. Notably, diagnosis did not moderate the relationship between IQ and whole hemisphere SA, indicating that this effect may have some specificity for visual cortex. In contrast, age-effects were not specific to visual cortex, and didn’t differ by group. These results suggest that our previous findings, implicating atypical functional connectivity of the pericalcarine cortex in intellectual functioning in autism, extend to the underlying anatomical morphology, including CT and SA, and that these differences may emerge before mid-childhood/adolescence.