Agenesis of the Corpus Callosum and the Autism Spectrum

Background: Agenesis of the corpus callosum (ACC), the congenital absence of the major interhemispheric commissure, occurs in 1:3,000 life births and can result in severe developmental impairment. Mildly affected patients can have social and communication deficits in the autism spectrum, although this has not been systematically addressed. Moreover, in many ACC cases other structures besides the corpus callosum are affected; how they influence clinical outcome is unclear. Objectives: We hypothesize that many ACC individuals have autistic traits. We secondarily hypothesize that the severity of these autistic traits correlate with MEG (magnetoencephalographic) measures of functional connectivity in non-callosal brain regions. Methods: Individuals from our ACC cohort were sent age grouped Autism Quotient (AQ) tests: Child (4-11 years), Adolescent (12-15 years), and Adult (16+ years). Child and Adolescent AQ tests were completed only by parents; Adult AQ tests were also completed by the ACC individual. The AQ is a standardized questionnaire developed and validated by Dr. Simon Baron-Cohen to assess autism traits in high functioning individuals. 409 AQ tests were mailed out and 101 tests were used in the final analysis: 44 children, 19 adolescents, and 38 adults. We also collected MEG resting state brain activity using a 275-channel biomagnetometer on Adult ACC individuals who visited our lab (n=10, IQ = 99.56 +/- 16.4). Neural sources were estimated using novel adaptive spatial filtering techniques and connectivity was estimated using imaginary coherence. Results: In children, AQ scores averaged 70.9 ± 20.2 compared to 41.7 ± 18.6 of the control children (p < 0.0001). The clinical cutoff score is 76 out of 141, and 40.9% ACC children exceeded the cutoff compared to 4.3% of the control child group (p < 0.0001). In adolescents, 26.3% exceeded the cutoff score compared to 0% in the control adolescent group (p = 0.001). In self-reporting Adults, 20% passed the cutoff compared to 2.3% in the control adult group (p = 0.0004). In contrast, when parents complete the assessment, 40% of ACC adults cross the threshold when using a comparative score. We tested the correlation between total Adult AQ scores and MEG resting state brain connectivity. We computed at each voxel the correlation (Pearson's r) between AQ scores and functional connectivity in a hypothesis independent way, and corrected for multiple comparisons using the false discovery rate (FDR) statistic. We found that the degree of connectivity in the right ventrolateral prefrontal cortex inversely correlated with the parent reported AQ score (r = -0.86; p<0.1; FDR corrected). When self-reporting scores were considered, however, no significant correlation was identified. Conclusions: Many ACC individuals, up to 10 fold more than control groups, exhibit social characteristics that exceed the AQ threshold screen for autism. Within the adult cohort, we find a correlation between the degree of autistic symptoms and underconnectivity within the right ventrolateral prefrontal cortex, a region implicated in social cognition. As deficits in frontal lobe connectivity and decreases in corpus callosum volume have been associated with autism, our findings strengthen these associations and provide a means to study this complex disorder.