Language Abilities and Traits of Autism Are Aetiologically Distinct: Evidence From a Community-Based Twin Study of 12-Year-Olds

Background:  Atypicalities in communication are considered to be one of the core features of autism, and often include delays in the development of spoken language. While numerous studies have explored language in autism (see Groen et al., 2008 for a review), there have been only two previous twin studies exploring the association between language and autism-related traits (e.g. Dworzynski et al., 2007, 2008).

Objectives:  The present study aimed to investigate, in depth, the genetic and environmental underpinnings of the association between traits of autism and language as assessed by four different measures of language ability when twins were aged 12-years.

Methods: Parents of ~5,000 twins participating in the Twins Early Development Study (TEDS) completed questionnaires assessing traits of autism in the twins when they were aged 12. The twins completed four online language tests: the Figurative Language and Making Inferences tests taken from the Test of Language Competence (Wiig et al., 1989), which are designed to assess semantic and pragmatic language respectively, the Test of Adolescent Language (TOAL; Hammill et al., 1994), which is a test of adolescent language proficiency, and a vocabulary test derived from the Wechsler Intelligence Scales for Children (Kaplan et al., 1999). Multivariate twin model fitting explored genetic and environmental contributions to each measure individually, and estimated the degree to which traits of autism and language abilities share aetiological influences with one another.

Results: There was a modest phenotypic correlation between traits of autism and the four language measures (r<0.20 in all cases [p<0.001]). The best fitting multivariate twin model was a model that estimated the proportion of additive genetic, shared environmental (environmental influences that create similarities within a twin pair), and nonshared environmental (environmental influences that make two twins in a pair growing up in the same family different from one another) influences on each trait and which provided separate estimates for males and females. While traits of autism displayed high heritability, all four language measures displayed more modest heritability, and a considerable degree of nonshared environmental influences. Traits of autism showed low to modest overlap in genetic influence with all four language measures; all genetic correlations were in the region of 0.03-0.32. Shared environmental overlap was moderate, and nonshared environmental overlap was very low across all four language measures.

Conclusions: These results suggest that aetiological influences on traits of autism and language abilities are largely independent of one another, meaning that it is likely that different genes and environments influence language abilities and traits of autism at age 12 years. These results have implications for molecular genetic endeavours, and may go some way to explaining the wide range of language skills evident in the ASD population.