Category Representation in Autism and Unaffected Siblings

Background:

Although autism spectrum disorders (ASD) are highly heritable no single genetic mechanism has yet been identified.  Neuroimaging studies have converged on functional impairments in regions of the brain dedicated to processing social stimuli, such as faces.  However, it is unclear which dysfunctions are necessary correlates of autism, which may indicate vulnerability, and which are epiphenomenal.

Objectives:

Unaffected siblings of children with autism did not develop an ASD, but share many genetic and environmental factors with their affected siblings, especially when compared to standard control groups. Therefore, they are an ideal population for investigating brain systems associated with increased vulnerability to an ASD. Our objective is to characterize differences and similarities in brain activation to several kinds of social and nonsocial stimuli.  Target groups include children with autism, unaffected siblings of children with autism, and typically developing children. 

Methods:

To date, 34 children and adolescents have participated in this fMRI study.  Children with autism (n = 12) were characterized by scores on the ADOS and ADI-R, the SRS, and the Vineland II.  ASD and other neurodevelopmental disorders were ruled out in a group of age and IQ matched unaffected siblings (n = 14). In addition, we targeted typically developing children (n = 8), who did not have siblings on the spectrum.  In a blocked design, participants viewed black and white exemplars of stimuli from several categories: neutral faces, houses, objects, letters, and Arabic numerals.  

Results:

Regions that preferentially responded to faces (left amygdala, left inferior frontal gyrus, & right fusiform gyrus), houses (bilateral parahippocampal gyrus), and letters (left middle temporal gyrus) were defined from the data of typically developing adults.  Among the child groups, no differences were found in response to houses in the parahippocampal gyri, or in response to letters in the middle temporal gyrus.  However, differential patters were observed in regions that preferentially responded to faces. Within the left amygdala, both the children with autism (ASD) and the unaffected siblings (UAS) exhibited hypoactivation relative to typically developing children (TDC). Within the right fusiform, only the children with ASD showed hypoactivation. Within the left inferior frontal gyrus, the UAS showed hyperactivation relative to TDC and ASD. 

Conclusions:

Uniform activation to houses and letters among the groups suggest that neither generalized cortical hypoactivation nor global deficits to visual information processing are part of the neuroendophenotype of autism. Consistent with previous neuroimaging studies of individuals with autism, we found reduced activation to faces in the amygdala and fusiform gyrus. Unaffected siblings also exhibited reduced activation to faces in the amygdala, but showed normal activation in the fusiform.  So although hypoactivation in the limbic system to social stimuli may be associated with autism, it is not sufficient to cause the disorder.  Interestingly, unaffected siblings showed increased activation, relative to typical controls, in the left inferior frontal gyrus. It is possible that increased activation in the left inferior frontal gyrus is either a buffer against susceptibility to autism, or is part of a system of compensatory mechanisms.