Network Analysis of Autistic Language Learners Along the Small World Spectrum

Recent network analyses of vocabulary growth revealed important relationships between the structure of the semantic environment and early vocabulary acquisition in non-autistic children. However, autistic children may be less likely to encode associated features of novel objects, suggesting divergent processes for acquiring semantic information about words. We examined the expressive vocabularies of 815 non-autistic and 163 autistic children (words produced: M_Autistic = 183.06, M_Non-autistic = 182.91). We estimated their trajectories of semantic development using network analyses. Network structure was based on child-oriented word associations. We analyzed networks according to indegree, average shortest path length, clustering coefficient, and small-world propensity (features holistically contributing to “small-world” network structure). Analyses revealed that autistic and non-autistic children are sensitive to the structure of their semantic environment. However, group differences were observed, with an early peak in the autistic group's clustering coefficient (how closely connected groups of words are), followed by a sharp decline. Moreover, across each network metric, we found that autistic children had reduced small-world structure relative to non-autistic toddlers. Thus, group differences indicate that, although autistic children are learning from their semantic environment, they may be processing their semantic environment differently, the language input to which they are exposed differs relative to non-autistic children, or a combination of the two.