Understanding the Changing Face of Autism: Determining Language Profiles of Children with ASD at Age Three Years

Background:  

Estimates in the literature have long referenced that 20-50% of children are nonverbal (Eigsti, 2011; Lord et al., 2004; Tager-Flusberg et al., 2005). However, these estimates are likely inaccurate given earlier identification and treatment of ASD, and changes in diagnostic criteria. Consequently, fine analysis of natural language, beyond standardized scores, is needed to further understand the language profile of toddlers with ASD. Research in this area is emerging (Tek et al., 2013), but further research with large samples and comparisons to other delay groups is needed. 

Objectives:  

This study aimed to provide a detailed characterization of language abilities of children with ASD at age three, when most children receive a final diagnosis, and provide a more modern estimate regarding the percentage of ASD children with functional language.

Methods:  

Participants were recruited from the general population largely via the 1-Year Well-Baby Check-Up Approach (Pierce et al., 2011) and included 109 with ASD, 60 typically developing (TD), and 42 with language delay (LD; note: 52% of the LD group had a transient diagnosis and did not show persistent delay). Data were analyzed from a ten-minute free-play interaction between the child and parent when children were between 30-42 months of age (M=34.95). Spoken language was assessed globally, and children with 15 words or fewer were categorized as minimally verbal (Paul et al., 2013; Tager-Flusberg, 2009). Partial-interval scoring was conducted to code features of the child’s and parent’s speech (see table 1). MANOVAs were conducted to compare differences between groups.

Results:  

Thirty-four children with ASD used fewer than 15 words during the assessment, and four used no words.  MANOVA analyses revealed significant effects between groups for many language features (see table 1). The largest effect sizes were seen for word approximations (F=39.523, p < .000, η² = .275), full words (F=54.353, p < .000, η² = .343), MLU (F=40.126, p < .000, η² = .278), child initiations (F=43.591, p < .000, η² = .295), child response (F=19.441, p < .000, η² = .157), and Wh- questions (F=26.061, p < .000, η²= .200).  The aforementioned variables, also showed significant differences between ASD and TD children, as well as ASD and LD. See Figure 1 for a sample of language features across groups. Children with a persistent language delay versus a transient delay will be further analyzed in the future.

Conclusions:  

Thirty-one percent of the children with ASD presented as minimally verbal, and only 3.7% used no words at all, suggesting that the idea that up to 50% of children with ASD are nonverbal is outdated. Children with ASD showed delays across variables and showed more impairment than LD and TD groups. Children with ASD and LD had similar usage of grammatical markings, although both were reduced compared to TD children. The social use of language was markedly different in children with ASD compared to LD and TD, as children with ASD also showed reduced initiations and responses to their parent, although their parents initiated to them just as frequently as other parents.