Nonshared Environmental Influences on Language Development: A Monozygotic Twin Differences Study

Background: Although linguistic impairment is no longer a diagnostic criterion for autism spectrum disorder, it remains a key area of focus and concern for many individuals and families. Given the recent focus on genetic influences on both autism and linguistic impairments more broadly, the study of environmental influences remains underdeveloped. Prior behavioral genetic work on language development has emphasized the role of both genetic and environmental influences, with the specific finding that nonshared (or person-specific) environmental (NSE) influences account for 5-90% of individual variation depending on age and the measures employed. Monozygotic (MZ) twin difference analyses provide a unique opportunity to identify NSE effects by examining the extent to which differences in specific environmental variables, such as diet, are associated with differences in language outcomes within MZ twin pairs who share approximately 100% of their DNA. The present study builds on prior behavioral genetic studies of language (DeThorne et al., 2012; 2016) and offers to the best of our knowledge, the first application of the MZ twin differences method to understanding NSE influence on children’s language development. Objectives: This work investigates NSE associations between specific environmental measures (i.e., gap between birth, weight at birth, extent of breast feeding, reading exposure, media viewing, and parental warmth) and language development at two time points (mean ages 7 and 12 years). Methods: Participants consist of approximately 215 MZ twin pairs from the Western Reserve Reading and Math Project (WRRMP; Petrill et al., 2006), a longitudinal study of reading development and related skills. The analysis focuses on data collected from home visit 2 (HV2) and HV7, during 1^st and 5^thgrade respectively. Each annual home visit included standardized measures, language samples, and parent questionnaires; 7-8% of the sample was receiving speech-language pathology services when they entered the study. Language measures were loaded into two latent factors: a Productive language factor and a Formal language factor, measured by language sampling and standardized tests respectively. An MZ differences method will be employed to examine associations between the NSE measures and the two language factors. Results: Based on previous research, we predict that differential birth weight and breastfeeding will be associated with differential language outcomes. Although previous literature has linked reading exposure, media viewing, and parental warmth with language outcomes, few studies have been conducted within the context of a genetically sensitive design. To our knowledge there has been no prior research examining the association between gap between twin birth and language outcomes; research has shown, however, the second-born twin is more likely to have more neonatal complications (e.g., respiratory distress syndrome) than the first-born twin. Subsequently, predictions are less clear. Consistent with prior work, we predict NSE relationships to be moderated by the form of assessment. Finally, we anticipate increased NSE associations at the extremes of discordance compared with the full unselected sample. Conclusions: The present study will elucidate potential NSE factors that may impact language development, measured by both language sampling and standardized tests. Future directions and clinical implications will be discussed.