Evidence of Inaccurate and Inefficient Visual Speech Perception in Children with Autism Spectrum Disorders

Background: Speech perception is a complex multisensory process wherein visual information complements the acoustic stream to facilitate comprehension of linguistic meaning and communicative intent (Massaro, 1998).  In typically developing (TD) individuals, visual cues speed the processing of speech information (van Wassenhove, Grant, & Poeppel, 2005) and improve the accuracy and efficiency of speech perception across listening conditions (Calvert, Brammer, & Iversen, 1998).   Although prior work has noted differences in detection, processing, and preference of auditory speech stimuli in ASD (Bebko, Weiss, Demark, & Gomez, 2006; Gervais et al., 2004; Klin, 1991), mounting evidence suggests that children with ASD display impairments in perception of visual speech (Iarocci, Rombough, Yager, Weeks, & Chua, 2010; Smith & Bennetto, 2007; Williams, Whiten, & Singh, 2004; Woynaroski et al., in progress).  Deficits in visual speech perception may contribute to the inefficient and inaccurate multisensory speech perception that has been observed in older children with ASD (Smith & Bennetto, 2007; Woynaroski, et al., in progress).

Objectives: This study examined speech perception in older children with ASD. Specific research questions included:

a)    Do children with ASD display deficits in identification of unisensory auditory or visual speech stimuli?

b)    Do children with ASD respond more slowly than TD peers to unisensory auditory or visual speech stimuli?

c)  Do children with ASD display atypical audiovisual speech perception in comparison to TD peers?

Methods: Groups included 8 to 17 year-old children with ASD (n=18) and TD controls (n=18) matched for mean age, sex, and IQ. Participants viewed video recordings of a young woman speaking four CV syllables at a natural rate and volume with a neutral facial expression in unisensory visual, unisensory auditory, and audiovisual conditions. Instructions were unbiased (i.e., report what the speaker said), and the response mode was non-verbal (button-press).  Stimulus presentation and recording of accuracy and reaction time data were managed by E-prime software.

Results: Children with ASD displayed impairments in their ability to identify visual speech stimuli when presented in a unisensory context.  In contrast, no differences were found for identification of auditory speech stimuli.  Additionally, the ASD group showed slower reaction times in response to unisensory visual stimuli, but not unisensory auditory stimuli, when compared with TD peers.  Finally, the ASD group showed impairments in multisensory speech perception and atypical patterns of multisensory gain wherein they derived relatively greater benefit from auditory versus visual cues compared to TD controls. 

Conclusions: Older children with ASD display inaccurate and inefficient perception of visual speech stimuli, but similar auditory speech perception skills, when compared with TD peers of similar age and cognitive ability.  Inordinate deficits in the processing of visual speech perception may account, at least in part, for atypical patterns of multisensory speech perception in this population, and relate to broader symptoms of ASD.  Findings are discussed in relation to previous work, prevalent theory, and future directions.