Page 183 - Demo

Embodied Facial Emotion Processing in Autism and Social Anxiety1816ResultsGroup Differences in Physiological Resonance, Facial Emotion Perception, and InteroceptionIn line with previous work, we observed a reduction in automatic facial mimicry responses to happy facial expressions, as well as a less accurate and confident emotion recognition, in individuals on the autism spectrum. More specifically, less relaxation of the corrugator in response to happy expressions was observed in the AS group compared to the NC group, p = .001, effect size r = 0.39. Furthermore, compared to the NC group, individuals on the autism spectrum were less accurate in the recognition of fearful expressions, p < .001, effect size r = 0.47, and of sadness expressions, p = .001, effect size r = 0.39. An overview of misclassifications to specific emotion categories can be found in Figure S1 in the Supplemental Materials. The ceiling performance for happy expressions in both the NC group and the SA group did not allow for group comparisons (100% accuracy rates). Individuals on the autism spectrum also reported reduced confidence in the recognition of all facial expressions compared to the NC group (psemotions < .001, range effect size r: 0.45 %u2013 0.48 and pneutral = .002, effect size r = 0.36). Lastly, the intensity of both angry and fearful expressions was perceived as less emotionally intense in the AS group compared to the NC group, p = .021, effect size r = 0.29 and p = .021, effect size r = 0.37 respectively (see Figure 1 and Table S2 in the Supplemental Material for an overview of all group differences). Additionally, both clinical groups reported to be less accurate in the perception of their interoceptive signals (SA group: p = .002, effect size r = 0.37; AS group: p < .001, effect size r = 0.42) while attending more strongly to them in everyday life (SA group: p < .001 effect size r = 0.45; AS group: p = .010, effect size r = 0.29).Facial Muscle Activity and the Perception of Negative Facial ExpressionsAgainst our expectations, we did neither observe a significant three-way interaction between Emotion category, Corrugator activity and Group, nor a significant two-way interaction between Corrugator activity and Group in the model with Emotion recognition accuracy as outcome (see Table S4 in the Supplemental Material for the model fit). Thus, there was no evidence for differences in the relation between activity over the corrugator muscle region (%u201cfrowning%u201d) and recognition of (specific) negative emotional expressions between