The Use of Prosodic and Syntactic Cues to Understand Intent in Discourse by Children with Autism Spectrum Disorders

Background: Atypicalities in prosodic expression are a hallmark clinical feature of autism spectrum disorders (ASD), but we are only beginning to understand differences in prosody comprehension that might characterize production differences.  One important function of prosody is to structure interchanges in discourse (e.g., indicating when a question is asked that necessitates a response).  Previous studies on the use of prosodic cues for discourse structure have suggested intact functioning in ASD.  Still, it is possible that individuals with ASD may be able to understand what is communicated by prosody, but may have a different processing strategy for utilizing prosody in conjunction with other cues to understand the intended meaning of an utterance.

Objectives: The current study examined the ability of individuals with ASD to use prosodic and syntactic cues to understand intent in discourse. We investigated their ability to make decisions in discourse when prosodic patterns and sentence structure provided either congruent or incongruent information about intent, as is common in natural speech.

Methods: Participants were 30 individuals with high-functioning ASD and 30 typically-developing peers between the ages of 11 and 19.  Groups were matched on chronological age, gender, Full Scale IQ, and receptive language. All participants had Full Scale IQ and receptive language scores greater than 80.  Participants were asked to make judgments on a series of sentences that indicated a speaker’s intent in discourse. In this task, participants were asked to identify whether or not someone was asking them a question, by pressing “yes” or “no,” but were given no further instruction.  Stimuli included utterances that indicated intent via prosodic cues (e.g., final-rise in intonation), and syntactic cues (e.g., subject-verb inversion).  When both prosodic and syntactic cues were present, they could be congruent (e.g., “Is she going to the store?” or “She is going to the store.”) or incongruent (e.g., “You are going to the store?” or “Are you going to the store.”). 

Results: We found a group by stimulus type interaction, p<.05, partial eta squared=.08.  Individuals with ASD were less likely than their typically developing peers to identify utterances as questions when the syntactic structure indicated a question but the prosody did not, p<.05, partial eta squared=.09, and were marginally more likely than peers to identify utterances as questions when the prosodic structure indicated a question but the sentence structure did not, p<.10, partial eta squared=.06.  There were no group differences when prosodic and syntactic structures were congruent.  A post hoc acoustic analysis of the stimuli revealed that typically developing peers were sensitive to subtle (but meaningful) acoustic differences in prosodic patterns in incongruent stimuli that did not seem to affect the interpretations of participants with ASD.

Conclusions: This study suggests that individuals with ASD utilize prosodic and syntactic cues to discourse differently than typically developing peers.  Participants with ASD gave preference to prosodic cues over syntactic cues in cases of incongruence.  Despite this preference, participants with ASD were less sensitive than typically developing peers to subtle acoustic differences in intonation patterns.