The cascading influence of multisensory processing on speech perception in autism

Abstract

It has been recently theorized that atypical sensory processing in autism relates to difficulties in social communication. Through a series of tasks concurrently assessing multisensory temporal processes, multisensory integration and speech perception in 76 children with and without autism, we provide the first behavioral evidence of such a link. Temporal processing abilities in children with autism contributed to impairments in speech perception. This relationship was significantly mediated by their abilities to integrate social information across auditory and visual modalities. These data describe the cascading impact of sensory abilities in autism, whereby temporal processing impacts multisensory information of social information, which, in turn, contributes to deficits in speech perception. These relationships were found to be specific to autism, specific to multisensory but not unisensory integration, and specific to the processing of social information.

Published estimates of group differences in multisensory integration are inflated

“…Using empirical data and simulations, we show that high within-group variability, coupled with routine use of small sample sizes, leads to a proliferation of inflated estimates of group differences in multisensory integration. Our results explain why recent studies have failed to replicate the large reported differences between cultures(15), genders (13), and children with developmental disorders (16). Studies of multisensory integration must increase sample sizes by an order of magnitude to produce accurate and reliable estimates of group differences.…”

“…We suggest that a more likely explanation than a variety of post hoc moderators is that small sample sizes inevitably result in high variability in population effect estimates, causing wild swings in their magnitude and sign across studies (29,30). This contention is reinforced by noting that the ASD study with the largest sample size (N = 76) also reported the smallest effect, a population difference of only 3% (16). Assuming that this large study is a more accurate measure of the true effect size, we can ask what the expected effect size will be for studies with a more typical sample size of N = 36 (mean across the 13 comparisons made in the ASD studies).…”

Atypical audiovisual temporal function in autism and schizophrenia: similar phenotype, different cause

“…Lastly, we employed a bootstrapping procedure to directly compare P ( C = 1), σ sensory noise , and μ ( C =2) values between the ASD and SZ groups. Importantly, it must be acknowledged that the explicit aim of the modeling work here was not to contrast ASD and SZ groups directly, but to ascribe their reported impaired multisensory temporal acuity relative to matched controls (Stevenson et al ., , ; Noel et al ., ,b,c,d; Zhou et al ., ) to particular principled variables. Indeed, the age and gender distribution of patients with ASD and SZ differ in the general population, and hence, matching these groups directly for age and gender would have supposed a deviation from naturalistic conditions.…”

“…More precisely, recent work has suggested that individuals with diagnoses of autism spectrum disorder (ASD; Foss‐Feig et al ., ; Kwakye et al ., ; Foxe et al ., ; Stevenson et al ., ; Noel et al ., , ,b) and schizophrenia (SZ; Foucher et al ., ; Martin et al ., ; Su et al ., ; Balz et al ., ; Stevenson et al ., ) possess atypically large TBWs, particularly for speech stimuli. Given that (multi)sensory integration is a fundamental building block in the construction of perceptual and cognitive representations, it has been postulated that alterations in multisensory temporal function may partially scaffold the higher‐order deficits present in the conditions (Russo et al ., ; Brandwein et al ., ; Woynaroski et al ., ; Baum et al ., ; Stevenson et al ., ; Noel et al ., ). While ASD and SZ share clinical features (Volkmar & Cohen, ), as conceptualized in the Research Domain Criteria framework (RDOC; Insel et al ., ), they are considered diagnostically distinct conditions.…”

Developmental Recovery of Impaired Multisensory Processing in Autism and the Cost of Switching Sensory Modality

“…Whereas multisensory processing clearly influences how we perceive most biological events, particularly in instances when sensory evidence is ambiguous (Sumby and Pollack, 1954;Ross et al, 2007;Stevenson and James, 2009;Crosse et al, 2016), individuals with autism spectrum disorder (ASD) often do not benefit from the availability of multisensory information to the same extent as their neurotypical (NT) peers (de Gelder et al, 1991;Smith and Bennetto, 2007;Silverman et al, 2010;Irwin et al, 2011;Bebko et al, 2014;Stevenson et al, 2014a;Stevenson et al, 2014b;Foxe et al, 2015). We and others have suggested that impaired multisensory processing in ASD contributes to some of the commonly associated phenotypes such as atypical responses to sensory stimulation, and may even have detrimental effects on higher-order processes such as social interaction and communication (Ayres and Tickle, 1980;Martineau et al, 1992;Iarocci and McDonald, 2006;Foxe and Molholm, 2009;Beker et al, 2017;Stevenson et al, 2017). bioRxiv March 21,2019 In previous work by our lab, we demonstrated that multisensory gain increases steadily over the course of development using both simple audiovisual (AV) cues (Brandwein et al, 2011) as well as AV speech stimuli (Ross et al, 2011).…”