From Emotion Production to Emotion Comprehension: An Emotional Training Proposal for Children with ASD

Background: According to the “direct-matching” hypothesis (Rizzolatti et al., 2001), there is a link between motion perception and execution and the comprehension of an observed action is underlay by the activation of the internal representation of the same action in the observer’s brain. Toddlers with ASD show signs of motor impairment - emotional expressions included - that often persists in adulthood. This atypical kinematic might lead to an abnormal perception of others’ actions. For instance, the motoric schema for expressing emotions in people with ASD could be different from the emotional expressions observed in other people and this mismatch could explaind the difficulties in comprehending other people's emotions.

Objectives: The aim of the present study was to investigate the link between motion perception and execution in ASD, by testing the effects of training the production of emotional expressions - both facial and bodily - on emotion comprehension. We hypothesized that training the emotion production would have favoured a reorganization of motoric schema for emotion expression and in turn improved the emotion recognition.

Methods: 17 children with ASD (age (M=5.38, SD=1.92); IQ (M=74.85, SD=18.9); ADOS (M=6.8, SD=2.18)) participated to one week of training. The daily sessions focused on joy and anger separately on alternate days. Children initially watched a cartoon clips with characters experiencing anger/joy; then they listened nursery rhymes about emotions and produced the corresponding expressions; finally, children imitated emotional expressions with the help of pictures.To assess the children’s comprehension of emotions, they underwent naming, comprehension, and match-to-sample tasks before and after the training, using pictures of facial and bodily expressions.

Results: Separated repeated measures ANOVAs were performed in each behavioural task. Results showed that the naming accuracy generally increased over time (F= 8.88; p=0.015; ges=0.079), and was greater for faces than for bodies (F=14.42; p=0.004; ges=0.278). Interestingly, only the naming accuracy for bodily expressions increased significantly (t=-2.447, p=0.037). Comprehension was higher for facial than bodily expressions (F=5.989; p=0.044; ges=0.022). The effect of Time on comprehension was marginally significant (F=3.595; p=0.1; ges=0.036), and the interaction between Time and emotion approached significance (F=4.395; p=0.074; ges=0.005), with increased accuracy for happiness over time. In the match-to-sample task, accuracy was higher for faces than for bodies (F=17.364; p=0.004; ges=0.199) and the effect of Time approached significance (F=4.276; p=0.077; ges=0.076).

Conclusions: The results show a tendency towards improvement in all the behavioural tasks after the training, suggesting that training the production of emotional expressions can foster the emotion comprehension. In general, emotion recognition resulted more accurate with facial than bodily expressions, suggesting that recognition of bodily expression is still developping in preschooler with ASD. In line with that, our results suggest a major effect of the proposed training on the recognition of bodily expressions. These finsings highlights the importance of including the body language in the emotional trainings for children with ASD.