Multisensory Integration and Autistic Traits

Background: Sensory processing issues are one of the most common complaints in ASD. One area of sensory difficulties that has been the focus of intense research in recent years is multisensory integration, or the ability to bind auditory and visual information into a single, unified perception. One consistent finding has been that multisensory integration is an area of difficulty in ASD when the sensory information contains social or linguistic information (e.g., audiovisual speech), but results have been less clear with simple, non-sociolinguistic stimuli. This study aims to address this current area of ambiguity.

Objectives: Our objective for this study is to determine whether multisensory integration abilities are related to traits and symptomatology commonly associated with ASD, specifically during the processing of non-sociolinguistic sensory information.

Methods: Sixty-five undergraduate students completed a behavioural audiovisual detection task, and completed a battery of questionnaires assessing ASD-related traits and symptomatology. During the behavioural component of the study, participants were presented with auditory tones and visual sinusoidal gratings (Gabor patches) embedded in auditory and visual white noise, respectively. Stimuli were presented either separately (audio or visual), or simultaneously in combination (audiovisual). Eighty trials of each modality were presented, as well as 240 catch trials (1/2 of all trials), in which no tone or grating was presented. Participants were asked to press the space bar as soon as they perceived either a tone or a grating. Multisensory enhancement (ME) was measured by comparing accuracy rates during audiovisual trials to the accuracy rate predicted by the unisensory conditions assuming independent processing:

pAV_acc = A_acc + V_acc – A_acc * V_acc.

Participants then completed measures of autistic traits in general (Autism Quotient, Broad Autism Phenotype Questionnaire), social competency and communication (Social Responsiveness Scale, Multidimensional Scale of Social Competency), repetitive behaviours (Repetitive Behaviours Questionnaire), and sensory processing (Sensory Profile 2, Sensory Perception Quotient).

Results: Mean accuracies for auditory, visual, and audiovisual conditions are presented in Figure 1A. Measures of multisensory enhancement for individuals were found in the normal range, from -.12 (no enhancement) to .02 (enhancement). Importantly, correlations between ME and all measures of ASD-related traits and symptomatology were non-significant, and did not approach significance, with R-squared values ranging from 0.001-0.03. Scatterplots associated with each questionnaire can be seen in Figures 1B-H.

Conclusions: These results suggest that there is no relationship between ME of simple, non-sociolinguistic sensory information and autistic traits and symptomatology. While multisensory integration issues are well established with sociolinguistic stimuli, these data suggest that these issues may be restricted to social or linguistic information. These data align well with studies of multisensory integration in ASD using different paradigms. For example, processing of the temporal characteristics of multisensory stimuli are impacted in autism when social or linguistic information is present, but not simple, non-sociolinguistic information. Importantly, the lack of any relationship between ME and ASD traits spanned a range of symptoms, including repetitive behaviours, social communication, and sensory issues, suggesting multisensory integration may be associated with autism symptomatology only when sociolinguistic information is present.