General and Specific Predictors of Understanding Tense/Aspect in Young Children with ASD

Background: Children with ASDs vary considerably in their usage of grammatical morphemes, with some showing consistent omissions in production while others omit only sporadically. Their comprehension of grammar (wh-questions, sentence frames) also demonstrates variability in performance. Consistency in production has correlated with concurrent vocabulary and/or general cognitive abilities (Eigsti et al., 2007; Fein et al., 1996; Roberts et al., 2004), while consistency in comprehension has correlated with earlier vocabulary, speed of sentence processing, and/or diagnosis/adaptive functioning (Goodwin et al., 2011; Naigles et al., 2009, in press).  The current study investigates the extent to which earlier predictors of tense/aspect comprehension are general (i.e., vocabulary, IQ) or specific (e.g., prior use of verb suffixes) in ASD.

Objectives: We compare children with ASD’s early language and cognition with their later understanding of tense/aspect markers.

Methods: Children (n=14) were tested every four months for two years.  At visit 1, children had a mean age of 33 months, had begun intensive ABA therapy, had language scores comparable to 20-month-old typical children, and a mean Mullen Visual Reception T-score of 38.3.  At Visit 5, the children first viewed the tense/aspect video; they averaged 49.6 months of age and produced on average 40% of the words on the CDI checklist.  Their mean age-equivalents on the Vineland scales were 32 months (Communication) and 37 months (Motor).  At visits 2 and 3, CDI scores were obtained for total words produced and usage of noun and verb grammatical inflections; moreover, 30-minute spontaneous speech samples yielded measures of MLU, noun and verb tokens, and verb suffixes. The tense/aspect video contrasted two familiar events, one presented as ongoing (a girl washing a dolly) and the other presented as completed (the girl finished washing the dolly).  During the baseline trial, both events were presented simultaneously with a non-directing audio (“She is on both screens!”). During the test trials, the audio presented the verb with the ‘-ed’ suffix (first block; e.g., “She washed the dolly”) and ‘-ing’ suffix (2^nd block; “She’s washing the dolly”).  Children’s eye movements were coded off-line.  Measures included (a) amount of looking (attention) to both screens, (b) preferential looking to the matching screen, and (c) latency of looking to the matching screen.

Results: Children with larger overall vocabularies at Visits 2 and 3 demonstrated overall longer attention to both screens during the Aspect task at Visit 5 (rs>.600, ps<.05).  Children who were faster to look at the matching screen, though, were those reported to produce more verb suffixes at visits 2 and 3 (rs <-.535, ps<.05).  And children who actually preferred the matching screen in the Aspect task were those who produced more verbs, more verb suffixes, and longer MLUs in spontaneous speech at visits 2 and 3 (rs>.565, ps<.044).  MullenVR scores at visit 1 did not predict performance on the Aspect task at visit 5.

Conclusions: Overall language predicts later overall attention during this comprehension task.  However, children’s speed and accuracy of comprehension were only significantly predicted by their prior use of the specifically relevant verbs and verb suffixes.