The McGurk Effect in Children with ASD: Examining Unisensory and Multisensory Responses to Speech Cues

Background: The ability to synthesize information across sensory modalities into unique multisensory percepts is necessary for understanding the world’s complexities, including for understanding speech.  In children with ASD, several studies have shown reduced susceptibility to the McGurk effect, a multisensory speech illusion; some have also shown deficits in lip-reading, the unimodal (i.e., visual) speech cue. According to the principle of inverse effectiveness, when unisensory stimuli are weak in eliciting responses, they are likely to induce increased multisensory responsiveness in combination with other spatially/temporally proximal sensory inputs. One would expect that, if lip-reading skills are deficient in ASD, enhanced integration of the auditory speech signal would occur.  It is important to understand the interplay between unisensory and multisensory speech perception to clarify the nature of (multi)sensory deficits that may impact conversational and social interchanges for individuals with ASD.

Objectives: To examine the accuracy with which children with ASD perceive auditory and visual speech input, as well as explore multisensory integration of audiovisual speech cues using the McGurk effect.
Methods: This study included 16 children with ASD and 17 children with typical development (TD), 8-17 years of age. Children with ASD and TD did not differ on age (ASD: mean=12.6yrs; TD: mean=12.1yrs), gender (ASD: 81.3%male; TD: 82.4%male), or IQ (ASD: mean=108.5; TD: mean=107.3), (all ps>0.50).  Children with ASD were administered the ADOS and parents were administered the ADI-R to confirm diagnoses; all met cutoff thresholds on both instruments.  Participants completed a McGurk task; on audio-visual trials, auditory presentation of the syllable “ba” fused with visual input of a woman mouthing the syllable “ga” typically produces an illusory percept of “da” or “tha”.   Unisensory visual trials (i.e., no sound, woman mouthing “ga”), as well as auditory trials (i.e., still face, sound clip of “ba”) were also presented.  Participants indicated perceived syllables via button-press response. Group comparisons were conducted comparing the percent of trials on which auditory-visual fusion (i.e., perceiving “da”/“tha”) was reported on auditory-visual trials and comparing perceptual accuracy on auditory- and visual-alone conditions.

Results: No group differences were found on perceptual accuracy for auditory-alone trials, t(31)=-1.158; p=.26.  Children with ASD were 65.3% (SD=31.2) accurate, while children with TD were 77.8% (SD=30.3) accurate. Children with ASD were less accurate (27.6%; SD=19.2) than children with TD (45.4%; SD=22.2) on visual-alone (i.e., lip-reading) trials, t(31)=-2.435; p=.02. No group differences were observed in susceptibility to the McGurk effect (ASD: 28.9%, SD=35.0; TD: 29.5%, SD=28.8), indexed by report of fusion syllables “tha” and/or “da”, t(31)=-.548; p=.59. 

Conclusions: This study found deficits in lip reading abilities in children with ASD, but no differences in auditory speech perception or in susceptibility to the McGurk effect.  These findings suggest intact multisensory integration in response to speech stimuli in children with ASD, for whom the visual speech cue is difficult to perceive alone. Results are in keeping with other findings of intact audio-visual integration of simple, non-speech stimuli.  Therefore, future studies should explore multisensory integration across stimulus complexity levels to clarify the nature of multisensory deficits in ASD.