Do Children with High-Functioning Autism Appropriately Mark Contrastive Stress in Speech to a Partner?

Background: Expressive prosody is often atypical in people high-functioning autism (McCann et al., 2007; Paul et al., 2005; Shriberg et al., 2001). It has been suggested that pragmatic prosody, which has a communicative rather than grammatical function, may be particularly impaired in autism due to its contextual nature. One instance of pragmatic prosody is contrastive stress, used to emphasize new or distinguishing information. The prominence of contrastive stress is created through an increase in syllable pitch, amplitude, and length relative to the surrounding sentence. Previous studies using perceptual ratings to evaluate contrastive stress marking report misplacement of stress by individuals with autism (Baltaxe, 1984; Fine et al., 1991).

Objectives: This study employed acoustic analyses of speech from an interactive communication task to examine the marking of contrastive stress in a naturalistic setting. The use of increased pitch, amplitude, and length were measured, as were the location of this marking and the location of the longest pause in the instruction.

Methods: Participants were 8- to 14-year-olds with HFA or typical development (TYP), matched on language level, age and gender. Fifteen HFA and 11 TYP children gave listeners instructions to select objects from a display. Instructions were of the form "Pick up the big cup" (where a small cup was also present in the display). Audio recordings were analyzed using PRAAT software (Boersma & Weenink, 2008) which allowed for automated calculations of pitch, amplitude, and length of instructions. The syllable containing the primary vowel in the adjective (e.g. li in “little”), which we expected to be marked for contrastive stress, was compared to the average pitch, amplitude, and length of the carrier phrase. An increase in these features suggests the use of contrastive stress. The syllables to the left and right of the target syllable were also analyzed to examine whether placement of stress differed between groups. Finally, the location of the longest pause in the utterance (before the determiner, adjective, or noun) was analyzed.

Results: Both groups demonstrated the use of contrastive stress by marking the adjective with increased pitch, amplitude, and length. However, typically-developing children used changes in amplitude to mark contrastive stress more reliably than children with HFA. Position analyses showed that both groups place marking on the expected syllable, rather than the syllable before or after. Pause location did not differ reliably between groups and generally came before the determiner. Finally, children with HFA who had higher levels of Performance IQ demonstrated greater use of contrastive stress marking. 

Conclusions: When adjectives were included in an instruction, as called for by the referential context, both groups marked contrastive stress in similar manner. However some participants with HFA did not produce distinguishing adjectives when necessary, and others produced adjectives with contrastive stress marking when this was irrelevant from their partner’s perspective.