Page 117 - Demo

Facial Mimicry and Metacognition in Facial Emotion Recognition1154levels. (C) The relationship between predicted accuracy in recognizing happy expressions and the corresponding filtered, baseline-corrected and z-scored zygomaticus activity depending on autistic trait levels. In the model fits, accuracy was coded binomial (0-1 values). For illustrative purposes, predicted accuracies for mean values as well as mean values +/- 1 SD of the continuous variables social anxiety traits and autistic traits are depicted. Shaded areas represent confidence intervals.DiscussionIn our study, we provided evidence that autistic traits and social anxiety traits are distinctly related to bottom-up (i.e., mimicry) and top-down (i.e., metacognition) components in emotion perception. Specifically, while individuals with higher social anxiety traits had significantly less confidence in their performance regarding all emotion categories, despite an unaltered actual emotion recognition performance, individuals with higher autistic traits were less accurate in the recognition of emotions, and in particular of fearful expressions. Furthermore, individuals with higher autistic traits seemed to be less able to calibrate their confidence judgments to their actual emotion recognition performance, as they displayed a poorer metacognitive sensitivity. Unexpectedly, we did not observe alterations in emotion-specific facial muscle mimicry with regard to either anxiety or autistic traits. Yet, we found indications that mimicry of frowning, indexed by corrugator activation, might facilitate the recognition of sad expressions, whereas mimicry of smiling, indexed by zygomaticus activation (and potentially relaxation of the corrugator), might support the recognition of happy expressions. Crucially, both links were less pronounced with higher autistic traits, while there was only weak evidence for a negative effect of social anxiety traits on the link between zygomaticus activity and accuracy in recognizing happy expressions.Contradicting our expectations, we did not find a negativity bias (i.e., an improved recognition of negative expressions) with higher social anxiety traits reflected in our main analysis on the recognition accuracy of the displayed facial expressions. When using relative accuracy scores as an outcome (see Table S20 in the Supplemental Material), however, a better recognition of angry facial expressions with higher social anxiety trait levels could be observed. Given that an improved recognition of negative expressions in SAD was also not consistently found in the literature (Bui et al., 2017), the effect seems unstable and additional factors might play a role. For example, a heightened sensitivity to negative expressions in social anxiety (disorder) might only occur under brief exposure times or when