Multisensory Speech Perception in High-Functioning Children with Autism Spectrum Disorders

Background: Accounts of unusual responses to sensory stimuli abound in the ASD literature (Iarocci and McDonald, 2006). Reports of sensory disturbance have motivated modern theories proposing children with ASD have difficulty integrating information to derive meaning from their experiences, potentially due to atypical temporal binding (Frith and Happé, 1994; Brock et al., 2002). Recent research has demonstrated atypical integration of simple multisensory (i.e., visual-auditory) stimuli over time in ASD (Foss-Feig et al., 2010; Kwakye et al., 2011). It remains unclear how aberrant responses to environmental stimuli and irregular integration of low-level sensory stimuli in ASD impact higher-level multisensory processes linked to language learning and social skills, such as speech perception.

Objectives: This study examined multisensory speech perception in high-functioning children with ASD. Specific research questions included:

a) Do children with ASD show reduced multisensory integration in response to incongruent “McGurk” stimuli?

b) Compared to controls, do children with ASD integrate incongruent speech stimuli differently over time?

c) Do children with ASD display deficits in unimodal or congruent audiovisual speech perception?

d) Does multisensory speech perception vary according to ASD symptom severity as measured by Autism Diagnostic Observation Schedule (ADOS)?

e) Is performance on speech perception tasks correlated with everyday responses to sensory stimuli as indexed by the Sensory Profile Questionnaire (SPCQ)?

Methods: Groups included 8-17 y.o. children with ASD (n=18) and TD (n=18) matched for mean age, sex, and IQ. Audiovisual incongruent “McGurk” syllables (visual /ga/ + auditory /ba/) were presented at seven stimulus onset asynchronies from 0-300ms to examine integration of audiovisual speech information over time. CV syllables associated with the McGurk task (auditory /ba/, visual /ga/, and percepts consistent with fusion - /da/ and /tha/) were presented in audiovisual congruent, unimodal-visual, and unimodal-auditory conditions to clarify multisensory and unisensory capabilities in this population. Correlational analyses were conducted to examine associations between performance on multisensory tasks, symptom severity as measured by ADOS scores, and atypical responses to sensory stimuli as indexed by SPCQ scores. Instructions were unbiased (e.g. report what the speaker said), and the response mode was non-verbal (button-press).  

Results: Children with ASD displayed deficits relative to TD in bimodal congruent speech perception. While unimodal auditory accuracy was similar, unimodal visual accuracy was significantly reduced for children with ASD relative to TD. Multisensory speech perception was strongly associated with unimodal-auditory performance, but less so with unimodal-visual performance in ASD. Additionally, perception of bimodal congruent stimuli correlated with ADOS Communication scores. No significant differences were seen in the frequency of illusory McGurk percepts between ASD and TD over time; however, post-hoc analyses revealed that perception of incongruent stimuli correlated with Auditory Processing Scores on the SPCQ for ASD.

Conclusions: Results suggest atypical patterns of multisensory speech perception in children with ASD. Associations with symptom severity and sensory profiles support larger links between speech perception, communication skills, and broader behavioral characteristics in ASD. Findings are discussed in relation to previous work, prevalent theory, and future directions. These outcomes may have important implications for academics and clinicians.