A Look At the Input: Relationships Between Parental Speech and Child Vocabulary in Autism and Typical Development

Background: There is extreme variability in vocabulary acquisition among children with autism spectrum disorders (ASD). While some children display clear delays, others achieve age-appropriate vocabularies (Luyster et al., 2007). Given less reliance on social input for language development, children with ASD may rely more on associative word learning mechanisms to acquire their vocabularies. This hypothesis has yet to be directly tested. Research with typically-developing children (TYP) demonstrates that variation in the statistics of parental speech influences child vocabulary; for example, word frequency (number of times a word is spoken; Goodman et al., 2008) and density of co-occurrence (number of different words with which a particular word co-occurs; Hills et al., 2010), are associated with the age of acquisition of those words. Few studies have explored how parental input influences children with ASD, nor in non-English languages.

Objectives: This study investigates the relationship between linguistic properties of parental speech input, specifically word types, word tokens, lexical diversity, MLU, word frequency (WF), density of co-occurrence (DCo), and child vocabulary. Based on studies that found no differences in parental input between parents of children with ASD and parents of TYP children (Siller & Sigman, 2002), we predict no differences between groups for measures of parental input. Consistent with studies of the influence on child vocabulary (Swensen et al., 2007), we hypothesize that in both groups, higher WF and DCo will predict a greater proportion of children having acquired those words. We also detail important cross-linguistic data in French and English.

Methods: The child-directed speech of 44 parents was investigated (ASD and TYP parent-child dyads n =22) Children were matched on receptive language ability. Dyads were from English-speaking (En; n = 13 per diagnostic group) or French-speaking (Fr; n = 9 per diagnostic group) families. Parents engaged in a 10-minute parent-child interaction with a standardized toy set. Child vocabulary was measured six months before and after the interaction using the MacArthur-Bates Communicative Developmental Inventories. Interactions were transcribed using CHAT (MacWhinney, 2000). Measures of word types, word tokens, lexical diversity, MLU, and WF were obtained in CLAN (MacWhinney, 2000). DCo will be calculated in R.

Results: Preliminary analyses of a sample of 14 ASD dyads and 14 TYP dyads (8 En and 6 Fr per group) show no significant differences between diagnostic groups for the parental speech properties of word types, word tokens, lexical diversity, and MLU. Analysis of DCo is in progress. Once all transcriptions are complete, logistic regressions will be used to examine the influence of WF and DCo on individual children’s acquisition of a particular word.

Conclusions: This is the first study to compare multiple properties of parental speech between parents of children with ASD and parents of TYP children and examine their influence on child vocabulary. As early intervention for children with ASD increasingly recruits parent involvement, understanding what input factors can be modified to create an optimal environment is important for language development. Understanding how children with ASD acquire words will allow intervention programs to better target language outcomes.