Does Multilingual Exposure Have an Effect on the Severity of Autistic Traits?

Background: Current research suggests multilingualism may have a significant impact on various aspects of autism. Recent evidence shows that multilingual autistic individuals demonstrate improved executive function and communicative competence (Gonzalez-Barrero & Nadig, 2016; 2017; Reetzke et al, 2015; Uljarevic et al, 2017). These are skills that can be a challenge for individuals with autism (Baron-Cohen et al, 1997; Happé, 1993; Hill, 2004). If multilingualism affects these areas of challenge, it is possible that it could interact with the way in which autistic traits are expressed.

Objectives: This study investigates the relationship between multilingualism and the prevalence of autistic symptomatology, as measured by the social-responsiveness scale (SRS II).

Methods: Data from the children in this sample were taken from three separate research projects. These children were separated into two cohorts, based on language background information. Parents were asked to complete the SRS-2, which is a questionnaire aimed at identifying autistic traits and quantifying its severity. Cohort 1: 45 multilingual autistic children and 314 monolingual autistic children (mean age = 5.4 years, sd ± 0.99) were tested across 12 different countries. Multilingualism was identified by the researcher working with each child and based on whether the child was exposed to more than one language in their home environment. Cohort 2: 22 monolingual and 30 multilingual autistic children (mean age = 12.9 years, sd ± 11.05) were tested in the UK. Parents of these children completed an extensive language background questionnaire, where they reported on the number of languages spoken, proficiency and frequency of use in each language.

Results: We ran independent t-tests to assess whether the average SRS scores differed between the monolingual and bilingual groups, where the bilingual group was defined as any child exposed to more than one language. For the first cohort, the mean SRS scores were 152.31 (32.21) for the bilingual children and 160.53 (28.21) for the monolingual children, this difference was not significant (t=1.794, p=0.074). For the second cohort, an independent t-test also showed no significant difference between the average SRS scores for bilinguals (mean score = 78.772) and monolinguals (mean score = 81.125) (t(38.334) = -0.793, p-value = 0.4328). Finally, we looked at whether proficiency and frequency of use in multiple languages among the bilingual group predicted differences in SRS scores in the second cohort. The regression models revealed that level of multilingualism significantly predicted differences in SRS scores, with higher levels of bilingual proficiency and frequency of use predicting lower SRS scores (F(1, 30) = 2.415, p = 0.004, > 0.01; r² = 0.22).

Conclusions: SRS scores were not significantly different between monolingual and bilingually exposed groups. However, bilingual proficiency and frequency predicted lower presentations of autistic traits on SRS. The findings from this project will help inform our knowledge of autism and how environmental factors, such as multilingualism, may influence its presentation.