Neural Representations Underlying the Observation of Interpersonal Socio-Affective Touch in Adults with and without Autism Spectrum Disorder

Background: Humans can easily grasp the affective meaning of socio-affective touch when observing other`s interactions. This ability may be linked to theory of mind (ToM) and somatosensory resonance. For example, when we see one person consoling another by giving a hug, we can understand the meaning of the hug and imagine its warmth. However, individuals with autism spectrum disorder (ASD) exhibit impairments in the use of nonverbal communication such as social and reciprocal touch, which gives them a huge disadvantage in social interactions. Despite the importance of touch in social communication and the pervasiveness of this aversion in ASD, surprisingly little is known about touch communication impairments in ASD, ultimately limiting diagnostic tools and therapeutic interventions.

Objectives: The present study aims to develop an experimental paradigm to address where and how socio-affective touch communication is represented in the brain of individuals with and without ASD. In particular, we examined the involvement of ToM and somatosensory resonance networks in processing observed social touch communication in the neurotypical (NT) brain in order to compare it with the brain processing of individuals with ASD.

Methods: A novel socio-affective touch database of 39 videos, covering pleasant (e.g., hugging a person), and unpleasant social touch (e.g. slapping a person), was used as stimuli (Figure 1). First, 21 NT participants evaluated valence and arousal of each video from which we extracted a group affective similarity matrix. Subsequently, the participants underwent fMRI scans during which they watched the same videos and performed an orthogonal task. Importantly, the participants also received positive and negative touch on their ventral forearms in the scanner in order for us to capture individual touch sensitive cortices as regions of interest (ROI). The social brain regions, including the temporoparietal junction (TPJ) known as a ToM area, were also included as ROIs. Using correlational multi-voxel pattern analysis (MVPA) methods, we generated a neural similarity matrix in each ROI. Lastly, multiple regression analysis was implemented in order to predict the neural data based on the perceived affect (combination of valence and arousal) of socio-affective touch.

Results: The results from pleasantness and arousal ratings suggest that positive touch was perceived as pleasant and negative touch as unpleasant with high inter-subject consistency on ratings. For example, the participants perceived a passionate hug as pleasant and highly arousing and a nudge as unpleasant and mildly arousing. Importantly, this affective meaning of touch was well represented in social brain areas and some parts of touch sensitive somatosensory regions. However, neural patterns in visual cortices as control regions could not be explained by perceived affect of observed touch communication (Figure 2).

Conclusions: Our findings highlight the involvement of perspective taking and somatosensory resonance when observing other`s affective touch interactions. Our results provide pivotal insights on neural mechanisms underlying the observation of interpersonal socio-affective touch in NT individuals. We are currently administering the same paradigm in 23 matched adults with ASD. At the conference we will be able to present group comparisons for the behavioral and neural data.