Is Atypical Multisensory Integration in Autism Specific to Socially Contingent Information?

Background: The ability to use multisensory integration (MSI) (i.e., simultaneously integrate information from multiple sensory modalities) allows us to interact adaptively and efficiently with our surroundings by creating a unified and coherent internal representation of the external environment. Both empirical evidence and anecdotal accounts suggest that a decreased ability to integrate information from different senses may partially underlie sensory-related behaviors in autism (Iarocci & McDonald, 2006; Donohue et al 2012). The majority of the evidence supporting altered MSI in autism stems from studies using visual and auditory stimuli that are socio-communicative in nature, such as speech or human faces (Magnée et al 2008; Silverman et al 2010; Smith & Bennetto, 2007). It has been suggested that MSI deficits in autism may be limited to social stimuli, and that this effect may not generalize to low-level sensory information void of social or linguistic characteristics (Bebko et al 2006; Magnee et al 2008; Mongillo et al 2008).

Objectives: To assess MSI in autism using an auditory guided visual illusion task based on sensory information that is non-social in nature (i.e., beeps and flashes) in order to determine whether the purported MSI deficits in autism are generalizable to non-social stimuli.    

Methods: Ten individuals diagnosed with ASD and ten individuals in a typically developing comparison group, matched for full-scale IQ, were asked to complete a computerized visual illusion task (Shams, Kamitani, & Shimojo, 2002) that assessed susceptibility to auditory-guided visual illusions. In this task, participants were required to determine whether they had perceived 1 or 2 flashes (F) while simultaneously hearing 0, 1, or 2 beeps (B). Participants were exposed to four non-illusion trials (i.e., 2F2B, 2F0B, 1F1B, 1F0B) and two illusion trials, whereby a discordant number of flashes and beeps were presented; (a) the fission illusion trial containing 1 flash and 2 beeps (1F2B), and (b) the fusion illusion trial containing 2 flashes and 1 beep (2F1B). Efficient MSI typically results in susceptibility to the illusion, with responses driven by number of beeps (B) presented (i.e., perceiving 2 flashes for the 1F2B fission trials). Susceptibility was measured for each group across illusion and non-illusion trials. 

Results: A mixed-model ANOVA was conducted to determine if group-differences in illusion susceptibility (i.e., decreased accuracy for identifying number of flashes) existed across experimental conditions. Results indicated a significant decrease in accuracy for the illusory compared to the non-illusory conditions for both autism and control groups (i.e., the illusions are present). However, significant group differences were not found for any condition.

Conclusions: Comparable between-group performance on the auditory-guided visual illusion task used in the present study indicates that individuals with autism are able to efficiently integrate low-level, visual and auditory information that is void of social content (i.e., beeps & flashes). This result is consistent with the suggestion that atypical MSI in autism may be specific to situations that call for integrating socially laden information, such as faces and voices.