Macrocephaly in autism pedigrees. C.K. Deutsch^1,2, S. Folstein^1,3, H. Tager-Flusberg^1,4, K.G. Gordon-Vaughn^1,3. 1) Psychobiology Program, Eunice Kennedy Shriver Ctr, Waltham, MA; 2) Harvard Medical School, Boston, MA; 3) New England Medical Center, Boston, MA; 4) University of Massachusetts, Boston, MA.

   A number of laboratories have found increased brain volume and macrocephaly to be statistically overrepresented in autism. Is this cranial enlargement uniform, or during development are some brain and cranial vault regions disproportionately affected? We have used objective, quantitative measures of craniofacial dysmorphology to assess cranial dimensions in a sample of 108 autistic probands. We have confirmed a statistical excess of increased head circumference in autism (25.0%; macrocephaly defined as a z-score greater than 1.5), exceeding the population baserate of 6.7% (p < .01). Goodness-of-fit tests reveal that the circumference scores in the autistic group are shifted in the direction of higher scores relative to the general population, yet a Gaussian distribution is still maintained. There was a marked increase in cranial width measured between the left and right eurion (p < .01). In contrast, the cranial length (the anterior-posterior linear distance between the glabella and opisthocranion) was normal, creating a brachycephalic pattern. Further, the presence of macrocephaly and increased biparietal width are predictive of the autistic probands' behavioral phenotype. The subgroup of patients with enlarged cranial dimensions is particularly intellectually impaired in comparison to autistic individuals with normal head size, with a specific impact on language skills. The dysmorphic phenotype described above for autistic probands was also seen among their first-degree relatives, even among those who did not meet criteria for autism (N = 50). The rate of macrocephaly among parents (N = 39) was 15.8% (p < .05) and among non-autistic siblings (N = 11) was 36.4% (p < .01). The pattern of excessive cranial width but not length was maintained among relatives (p < .05 and n.s., respectively). These data imply a specificity of brain growth patterns in autistic macrocephaly that is conserved among relatives. The implications of these findings for models of brain maldevelopment are also discussed. (This research was supported by NIH PO1 DC03610, R01 MH55135, and P30 HD04147.).