Page 99 - Demo

Facial Mimicry and Metacognition in Facial Emotion Recognition974levels were further expected to be associated with less facial mimicry, and this reduction was expected to be strongest for negative expressions. Furthermore, as automatic facial mimicry has been suggested to be impaired in ASD, the information about facial muscle activity might also be less well integrated in emotion recognition. Accordingly, we explored whether a positive relationship between facial mimicry and emotion recognition would be less pronounced in individuals with higher autistic trait levels. Lastly, extending on the few findings in clinical samples, we expected lower metacognitive sensitivity in relation to higher autistic traits. Hence, confidence judgments should be less predictive of actual emotion recognition accuracy in individuals with higher autistic trait levels. Given the little and inconclusive evidence on metacognition in emotion recognition in healthy and clinical populations, our analyses regarding this research question were explorative.MethodsParticipantsFifty-seven healthy participants were recruited from the Leiden University student population (50 female and seven male). Their ages ranged from 18 to 30 years old (M = 22.75, SD = 3.27) and they all reported normal or corrected-to-normal vision. None of the participants reported current or past psychological or neurological disorders. Participation in the study was voluntary and written consent was obtained prior to the experiment. Participants received either two university credits or a monetary reward of six euros as reimbursement. The study has been executed in accordance with the Declaration of Helsinki and approved by the local ethics committee of the Faculty of Social and Behavioral Sciences at Leiden University (# 2020-02-10-M.E. Kret-V1-2117). In the scope of a Master thesis project, an a priori power analysis was run for this study, treating clinical traits as a categorical variable (low vs. high trait levels). Based on a similar previous study that found significant group effects with medium effect sizes (Zwick & Wolkenstein, 2017), we estimated our ideal sample size with the Power Analysis for General ANOVA application (PANGEA) (Westfall, 2015). With 30 participants per group, hence 60 participants in total, a group effect of d = .50 should be found with a power of .901. Because of the COVID-19 regulations