19704
Form Is Easy, Meaning Is Hard: What Language Comprehension Reveals about Language in Autism
Children with ASD demonstrate consistent delays in the onset of language development; many also show slower growth rates in expressive language (1,8). However, given their social impairments, their comprehension of language has been more difficult to ascertain because participation on conventional receptive language tasks can be challenging (2). New methods, such as Intermodal Preferential Looking (IPL), show promise for illuminating language comprehension in children with ASD. We have used IPL to investigate comprehension of several aspects of grammar, and several word learning strategies, in a longitudinal study of children with ASD (5).
Objectives:
We bring together findings from 6 IPL studies of language comprehension and discuss what they reveal about the strengths and weaknesses of early development in ASD.
Methods:
Children were assessed every four months for a total of 6 visits. At the onset of the study, 17 children with ASD (MA=33 months) and 18 TD control children (MA=20 months) were matched on CDI vocabulary production scores (M-TD=118.7 words, M-ASD=94.7 words). Children viewed IPL tasks at each visit (see Table 1); two assessed word learning (NounBias: Do children map novel words onto objects rather than actions; ShapeBias: Do children extend novel words to objects of the same shape), three assessed grammatical comprehension (WordOrder: Do children understand sentences in SVO order; Wh-questions: After children are shown ‘hitting’ events (e.g., an apple hitting a flower), do they understand that ‘What did the apple hit?’ refers to the flower whereas “What hit the flower?” refers to the apple; Aspect: Do children understand that –ing is used for ongoing events and –ed for completed events. The final video assessed children’s ability to use transitive frames to learn causative verbs (Syntactic Bootstrapping). Children’s eye movements were coded off-line. For all tasks, children should look longer and more quickly to the matching screen during the test trials compared with the control trials. Production measures were also obtained, from parent-child play sessions at each visit.
Results: The grammatical tasks yielded similar findings in TD and ASD groups: The children with ASD showed early consistent comprehension of SVO word order, and reliable but developmentally-delayed comprehension of wh-questions (3,6). Both structures were observed in children’s comprehension earlier than in production. Good comprehension of the ‘-ing/-ed’ aspectual contrast was observed, even among low-verbal children (9). Performance correlated with joint attention skills, although initial language levels accounted for more variance. The lexical tasks yielded mixed findings: In both groups, novel words presented in isolation were mapped onto objects rather than actions; moreover, novel verbs in transitive sentences were consistently mapped onto causative over noncausative actions (4,6,7). However, one word learning task revealed serious impairments in the ASD group, who never demonstrated a consistent Shape Bias (7).
Conclusions:
Whereas grammatical comprehension was relatively robust in the ASD group, one word learning strategy was found to be significantly impaired. Making semantic generalizations, especially those requiring attention to global visual properties, seems more challenging than making syntactic generalizations. Social factors played only a small role in explaining variance in comprehension performance.