Intact neural representations of affective meaning of touch but lack of embodied resonance in autism: a multi-voxel pattern analysis study

Abstract

Background: Humans can easily grasp the affective meaning of touch when observing social interactions. Several neural systems support this ability, including the theory of mind (ToM) network and the somatosensory system linked to embodied resonance, but it is unclear how these systems are affected in autism spectrum disorder (ASD). Individuals with ASD exhibit impairments in the use of nonverbal communication such as social and reciprocal touch. Despite the importance of touch in social communication and the reported touch aversion in ASD, surprisingly little is known about the neural systems underlying impairments in touch communication in ASD. Methods: The present study applies a dynamic and socially meaningful stimulus set combined with functional magnetic resonance imaging (fMRI) to pinpoint atypicalities in the neural circuitry underlying socio-affective touch observation in adults with ASD. Twenty-one adults with ASD and 21 matched neurotypical adults evaluated the valence and arousal of 75 video fragments displaying touch interactions. Subsequently, they underwent fMRI while watching the same videos. Using multi-voxel pattern analysis (MVPA) and multiple regression analysis, we examined which brain regions represent the socio-affective meaning of observed touch. To further understand the brain-behavior relationship, we correlated the strength of affective representations in the somatosensory cortex with individuals' attitude towards social touch in general and with a quantitative index of autism traits as measured by the Social Responsiveness Scale. Results: Results revealed that the affective meaning of touch was well represented in the temporoparietal junction, a core mentalizing area, in both groups. Conversely, only the neurotypical group represented affective touch in the somatosensory cortex, a region involved in self-experienced touch. Lastly, irrespective of the group, individuals with a more positive attitude towards receiving, witnessing, and providing social touch and with a higher score on social responsivity showed more differentiated representations of the affective meaning of touch in these somatosensory areas. Conclusions: Together, our findings imply that male adults with ASD show intact cognitive understanding (i.e., "knowing") of observed socio-affective touch interactions, but lack of spontaneous embodied resonance (i.e., "feeling").