Are Autistic Traits Associated with Compromised Audiovisual Integration of Socially Relevant Information?

Background:

This study investigated the relationship between autistic traits and the perception of audiovisual displays of human motion. Previous research has demonstrated that, in typical observers, visual sensitivity to point-light walkers improves when paired with footstep sounds, but not when paired with meaningfully unrelated sounds. Superior temporal sulcus and premotor cortex are implicated in biological motion perception (Saygin, 2007) and audiovisual integration (Barraclough et al., 2005; Keysers et al., 2003). These regions appear to be compromised in individuals on the autism spectrum (Pelphrey & Carter, 2008; Oberman et al., 2005). Thus, compromised audiovisual integration of human actions may be associated with autism. Impairments in the integration of such socially relevant information could lead to difficulties in social interaction. There is no consensus as to whether multisensory integration is compromised in individuals with Autism Spectrum Disorder (ASD).  Some studies report intact audiovisual processing (Keane et al., 2010) while others report deficiencies (Smith & Bennetto, 2007). To address this inconsistency, multisensory integration was tested in typical individuals as a function of the magnitude of their autistic traits. Typically developed adults vary in the degree to which they possess autistic traits (Baron-Cohen et al., 2001). Here, we investigated the relationship between autistic traits, as measured by the Autism Quotient (AQ) and sensitivity to audiovisual displays of human movement.

Objectives:

The aim of this study was to determine whether autistic traits are associated with the integration of auditory and visual cues to human actions in typically developed adults.  

Methods:

Eighty-three typical adults completed the AQ and an audiovisual biological motion detection task. Point-light displays of human walking motion (3000ms duration) were presented coherently (person present) or scrambled (no person), and hidden within a point-light mask. In a between-subjects design, participants heard either footsteps (meaningfully related) or frequency-scrambled footsteps (not meaningfully related) that were coincident with the footfalls of the point-light walker. On each trial, participants indicated whether or not a walker was present.

Results:  

AQ scores ranged between 9 and 31 (mean 18.83). Separate correlational analyses were conducted for the coherent footstep and the scrambled footstep groups. Analyses revealed a significant negative correlation between task accuracy (% correct) and AQ score for the footstep group (r = -.373, p = .021) but not for the scrambled footstep group (r = -.121, p = .430). Thus, as autistic traits increase, sensitivity to meaningfully related audiovisual biological motion cues decreases, while sensitivity to meaningfully unrelated audiovisual biological motion cues does not.

Conclusions:  

In the real world, social interactions include both visual and auditory cues. Social capabilities are constrained by one’s ability to integrate cues that are meaningfully related and segregate cues that are meaningfully unrelated. We observed a negative correlation between autistic traits in typical observers and the ability to detect human motion in the presence of meaningfully related footstep sounds. This supports the hypothesis that the integration of meaningfully related auditory and visual biological motion cues is impaired increasingly with increases in the magnitude of autistic traits.