Agenesis of the corpus callosum and autism: a comprehensive comparison

Abstract

The corpus callosum, with its ∼200 million axons, remains enigmatic in its contribution to cognition and behaviour. Agenesis of the corpus callosum is a congenital condition in which the corpus callosum fails to develop; such individuals exhibit localized deficits in non-literal language comprehension, humour, theory of mind and social reasoning. These findings together with parent reports suggest that behavioural and cognitive impairments in subjects with callosal agenesis may overlap with the profile of autism spectrum disorders, particularly with respect to impairments in social interaction and communication. To provide a comprehensive test of this hypothesis, we directly compared a group of 26 adults with callosal agenesis to a group of 28 adults with a diagnosis of autism spectrum disorder but no neurological abnormality. All participants had full-scale intelligence quotient scores >78 and groups were matched on age, handedness, and gender ratio. Using the Autism Diagnostic Observation Schedule together with current clinical presentation to assess autistic symptomatology, we found that 8/26 (about a third) of agenesis subjects presented with autism. However, more formal diagnosis additionally involving recollective parent-report measures regarding childhood behaviour showed that only 3/22 met complete formal criteria for an autism spectrum disorder (parent reports were unavailable for four subjects). We found no relationship between intelligence quotient and autism symptomatology in callosal agenesis, nor evidence that the presence of any residual corpus callosum differentiated those who exhibited current autism spectrum symptoms from those who did not. Relative to the autism spectrum comparison group, parent ratings of childhood behaviour indicated children with agenesis were less likely to meet diagnostic criteria for autism, even for those who met autism spectrum criteria as adults, and even though there was no group difference in parent report of current behaviours. The findings suggest two broad conclusions. First, they support the hypothesis that congenital disruption of the corpus callosum constitutes a major risk factor for developing autism. Second, they quantify specific features that distinguish autistic behaviour associated with callosal agenesis from autism more generally. Taken together, these two findings also leverage specific questions for future investigation: what are the distal causes (genetic and environmental) determining both callosal agenesis and its autistic features, and what are the proximal mechanisms by which absence of the callosum might generate autistic symptomatology?

Keywords: autism; connectivity; corpus callosum; developmental neuropathology; social cognition.

Figure 1

Structural MRI of participants with agenesis of the corpus callosum. Here we divided the entire AgCC group, based on neurological criteria, into those with complete agenesis and those with partial agenesis, and for each constructed a midspace template by iterative non-linear registration to an initial MNI-space target template of all T₁-weighted structural images from that group (Tyszka et al., 2011): (A) midspace template for complete AgCC (n = 21); (B) representative anatomy in a participant with complete AgCC; (C) midspace template for partial AgCC (n = 5); and (D) representative anatomy in a participant with partial AgCC. Diffusion studies indicate that structural connectivity of the remnant corpus callosum in partial AgCC can be highly variable (Wahl et al., 2009). Therefore, the location of residual callosum cannot be assumed to represent the same connectivity as that region within an intact corpus callosum.

Figure 2

Effect sizes (partial η²) from between-group comparisons on (A) diagnostic measures and (B) measures of related social skills. Non-significant findings are shown in dark blue. In A measures on the y-axis are listed in order of greatest to smallest effect size in comparisons of AgCC⁻ and ASD. ***P < 0.001, **P < 0.01, *P < 0.05 in three-subgroup univariate ANCOVA.

Figure 3

ADOS algorithm scores. Individual participants’ data are shown on the (A) communication, (B) social, (C) total (communication + social) and (D) restricted behaviour scales. Higher scores indicate greater symptomatology, with black dotted line denoting the cut-off for autism diagnosis criteria and gray line the cut-off for autism spectrum criteria. Circles represent individual participants, with subgroup indicated by colour (AgCC⁺ = blue, AgCC⁻ = red, ASD = green) and x-axis location (jittered to show individual circles more clearly). ASD diagnosis on the ADOS requires scoring above the cut-off on all three of the algorithm scales: communication, social and total.

Figure 4

Parent ratings on autism diagnostic scales. Individual scores for the parent rating scales including (A) SCQ raw score (B) ADI-R communication, (C) ADI-R social, and (D) ADI-R restricted behaviour algorithms. Higher scores indicate greater symptomatology. On ADI-R scales, the black dotted line indicates lower-limit of autism spectrum criteria.