First degree relatives of autistic individuals are reported to have high rates of macrocephaly (6
). In addition, it is anticipated that approximately 20% of HR infants in this sample will meet criteria for an autism spectrum disorder by 36 months of age (17
). In this first report of brain volume in 6 month old infants at high familial risk for autism from an ongoing longitudinal neuroimaging study, we find no differences in overall brain tissue volumes, head circumference, or ventricular volumes compared to LR peers.
This investigation does not address the question as to whether early brain volume differences can be observed in those HR children who are later diagnosed with an ASD, which will be examined in the longitudinal dataset from this study. Rather our intent was to study whether differences in brain volume could be detected in a sample of children at high genetic risk for autism, an approach that has been used previously to examine familial rates of macrocephaly in ASD (7
). There is, therefore, a chance that using outcome data we will be able to detect significant differences at 6 months once we are able to examine trajectories of growth in the various subgroups at 24 months of age.
The absence of brain volume differences in six month olds at high familial risk for autism in this large cohort is consistent with our hypothesis that brain enlargement in autism as well as possibly in those with high genetic liability is a later occurring post-natal process. Disruptions of cortical maturation related to impaired experience-dependent synaptic development have been observed in mouse models of both Angelman Syndrome (28
) and Fragile X syndrome (29
), neurogenetic developmental disorders that have behavioral features which overlap with ASDs. There is also the report of mutations associated with the defective expression of activity-driven genes in individuals with ASDs (30
), although it is unclear whether this mechanism accounts for more than a very small proportion of genetic variance underlying ASDs. Other hypotheses about brain overgrowth, based on the observation that early brain overgrowth is associated with increases in cortical surface area (10
), suggest that brain overgrowth in the latter part of the first year may be the result of an increase in neuronal precursors, perhaps related to aberrant molecular regulatory processes (31
). The findings from the current study suggest that early postnatal events may underlie latter occurring brain overgrowth and raise the optimistic possibility that there is a window of opportunity where early postnatal intervention, during a period of tremendous brain plasticity, may have an important impact on later emergence of autistic behavior.
Studies examining early behavioral markers of autism have shown infants who later go on to receive a diagnosis of an ASDs do not exhibit overt autistic behaviors at 6 months of age (15
). However, by 12 months of age, a time when we anticipate brain differences may begin to emerge, behavioral deficits can be detected in those infants who are later diagnosed with autism (15
). Work by our group examining the behavioral characteristics of our sample is underway. In the current study, the HR group likely includes children who later receive a diagnosis of autism. With the later addition of diagnostic outcome data on this sample, it will be possible to examine whether brain volumes for those children who are HR and eventually manifest ASDs differ from those HR children who do not go on to develop ASDs, to specifically test our hypothesis of brain overgrowth in the latter part of the first year of life or early second year of life.
We also did not detect any clinically significant radiologic findings specific to the HR group. The types and rate of incidental findings observed was the same between groups, as was the degree of enlargement for the subarachnoid spaces and the periventricular spaces. Most reports looking for evidence of clinical neuroradiologic abnormalities in individuals with autism have not detected a significant type or pattern of findings associated with the disorder. Scattered reports of individuals with neuronal migration abnormalities (33
), as well as subcortical abnormalities in the corpus callosum (34
) and increased rates of dilated Virchow-Robin spaces (35
) exist. However, numerous methodological shortcomings such as the etiologic heterogeneity known to be inherent in autism, small sample sizes, variations in subject age and inconsistent findings across studies have complicated interpretation of this research. One large-scale retrospective study (36
) observed elevated rates clinically reported abnormalities in children with autism compared to medical patients. Forty-eight percent of the ASD cases were rated as having abnormalities. The primary abnormalities observed included white matter signal intensity abnormalities, dilated Virchow-Robin spaces, and temporal lobe abnormalities (e.g., subcortical hyperintensities in the temporal poles). White matter abnormalities (e.g., punctate or posterior T2 hyperintensities) were present in younger children (e.g., 5 year olds versus 7 year olds), while the temporal lobe abnormalities showed no developmental trend. It should be noted that the current study differs from this report by Boddaert (36
) as we do not explicitly examine children (2 to 16 years of age) with ASDs, but rather we examine infants at high risk for autism. The developmental nature of incidental findings we observed and their association with ASDs therefore remains a question worthy of future investigation.
A limitation of this study, due in part to our focus on six month olds, is that, due to the difficulty associated with segmenting gray and white matter at this early age in development, we are unable to define regionally-specific brain volumes. White matter structure and myelination may best be captured with other methods, such as diffusion tensor imaging (DTI), and we have recently reported significant differences in white matter development in HR infants who later receive a diagnosis of ASD (37
). More dynamic methods like resting state functional MRI (fcMRI) may also detect early functional changes prior to the occurrence of more grossly observable changes in volume. These methodologies may be more sensitive to early brain differences and may be able to detect changes in structural or functional development that are not possible using standard tissue classification methods. Lastly, our current study is cross-sectional and at this point our sample is characterized in terms of HR and LR status. Only a portion of our current HR group may later be diagnosed with an ASD. We will include diagnostic outcome data and longitudinal imaging data from later timepoints (e.g., 12, 24 and/or 36 months of age) to examine the groups for more specific insights into the behavioral and biological mechanisms underlying the development of ASDs.