Poster #32 - Can expressive language features indicate whether low-verbal children with ASD will progress linguistically?

• In Event: Poster Session 11

Abstract

Significant heterogeneity exists in the language trajectories of children with autism spectrum disorder (ASD; Pickles et al., 2014). While many children with ASD move beyond the single-word stage to combine words and eventually achieve complex syntax, nearly 20% remain low/minimally verbal into adulthood (Bal et al., 2016), continuing to produce mainly single-word utterances and occasional word combinations. Prior work shows that receptive language scores and production of sounds and words can differentiate groups of autistic children with better and poorer overall language outcomes (Paul et al., 2008). However, these measures were based on standardized assessments and parent report data. Analysis of children’s speech during naturalistic language samples may identify additional linguistic variables that could be more relevant for morphosyntactic outcomes specifically.

In this project, we investigate the following question: Among children with ASD who begin at a low level of morphosyntactic development, can expressive language features differentiate those who progress linguistically from those who remain static?

12 children with ASD and six typically developing (TD) children were selected from a larger longitudinal project (Naigles & Fein, 2017). Language samples during parent-child play were collected and analyzed at two timepoints (Time1[T1] mean age=31.81 months; Time2[T2] mean age=51.70 months). At T1, no children regularly combined words (mean length of utterance [MLU]<2, word combinations <20% of utterances). At T2, six children in the ASD group still did not regularly combine words (“static-ASD”), while the other six in the ASD group (“change-ASD”) and all six in the TD group did regularly combine words (MLU>2, word combinations 40-55% of utterances). Group differences at T1 in expressive language (number of different words [NDW], proportion of word combinations, proportion of verb types, verb uniqueness) and non-expressive measures (receptive language scores, duration of response to joint attention [RJA] episodes) were qualitatively analyzed.

At T1, the static-ASD, change-ASD, and TD groups did not differ on NDW or proportion of word combinations. The change-ASD and TD groups generally produced a lower proportion of verb types than the static-ASD group, despite significant variability in all three groups (Figure 1a). Further, the change-ASD produced 19 unique verbs (verbs not produced by the static-ASD group), while the static-ASD group only produced six (Figure 1b). The change-ASD and TD groups had higher receptive language scores and engaged in more RJA at T1 than the static-ASD group (Figure 2).

In this small sample, broad measures of early expressive language (NDW, proportion of word combinations) did not differentiate a group of children with ASD who remain low/minimally verbal from a group of children with ASD who achieve phrase-level speech. More specific expressive measures (proportion of verb types, verb uniqueness) may better distinguish these two groups. The clearest differences appeared in receptive language and amount of RJA, consistent with previous work that suggests the importance of these variables in differentiating autistic children with better expressive language outcomes from those with poorer expressive language outcomes (Ellis Weismer & Kover, 2015; Paul et al., 2008). These results have implications for predicting language progress in young low-verbal children with ASD.

Authors

• Kaya J. LeGrand, University of Connecticut
  Presenting Author

• Letitia R Naigles, University of Connecticut