Page 184 - Demo

Chapter 6182either clinical group and the NC group, Figure 1. Group differences in physiological measures and facial emotion perception judgments. Graphs depict differences on subject-level averaged data by emotion category in (A) baseline-corrected, z-scored corrugator activity, (B) baseline-corrected, z-scored zygomaticus activity, (C) baseline-corrected, logtransformed maximum deflection in skin conductance, (D) emotion recognition accuracy, (E) confidence in emotion recognition, and (F) perceived emotional intensity. Significant group differences in Wilcoxon rank sum tests are indicated by asterisks, according to the significance levels: *** p < .001; ** p < .01, * p < .05. with Bayes factors below 0.3 supporting the null hypothesis, apart from the NC-SA difference for fearful expression (0.48). Yet, next to main effects of Emotion categoryand Group, we observed significant interactions between Emotion category and Corrugator activity, %u03c72 = 9.62, p = .022, as well as between Emotion category and Group, %u03c72 = 46.57, p < .001. In line with the results on the subject-level averaged data, recognition of sadness and fear displays was significantly less accurate in the AS group compared to the control group, ORSadness = 0.40, 95%CI [0.23 ,0.71], p < .001 and ORFear = 0.15, 95%CI [0.08, 0.30], p = .001. We additionally observed worse recognition of angry facial expressions in the AS compared to the NC group, OR= 0.43, 95%CI [0.22, 0.85], p = .011. Similarly, we only found group differences in rating both Confidence in emotion recognition and Perceived emotional intensity, independent of Corrugator activity, for the different levels of Emotion categoryF(6,3275) = 5.32, p <.001 and F(6,3275) = 4.69, p <.001 respectively (see Tables S5 and S6 in the Supplemental Material for the model fits). Looking at the Bayes factors, there indeed seemed to be evidence for no differences in the association