Page 81 - Demo

Physiological Resonance and Interpretation of Emotional Expressions793Subtle facial cues. Both linear and cubic polynomials significantly predicted changes in SKT in the subtle facial cue model, Flinear (1, 141599) = 7.225, p = .007; Fcubic (1, 141599) = 5.227, p = .022. Additionally, all interactions between the emotion category and the three polynomials were significant , Flinear*category (3, 141599) = 5.543, p = 0.001; Fquadratic*category(3, 141599) = 24.200, p < .001; Fcubic*category (3, 141599) = 5.095, p = 0.002. Thus, adding subtle emotional cues to a neutral picture might already make a difference in the characteristics of SKT changes in the observer. Consulting the model statistics (Table 6 in Online Resource 3) and the predicted value graph (Fig. 5c), both faces with added tears and faces with added dilated pupils were associated with an initial dip. While this dip turned into an increase after approximately 2s for the first (reaching a similar temperature level as the faces without cue), it did not for the latter. Faces with a blush yielded a strong increase in cheek temperature which attenuated over time. The subsequent bootstrap analysis did not support the directionality of any of the effects.Facial EMG Corrugator supercilii. The split-half tests on differences in facial muscle activity between emotional and neutral expressions within distinct time bins yielded emotion- as well as time-specific findings. When viewing happy compared to neutral facial expressions, activity over the corrugator supercilii region was significantly reduced in both our training and test sample 500ms, 600ms, 1600ms and 3800ms after stimulus onset (all ps < 0.05). Further, while 200ms and 600ms after stimulus onset, angry facial expressions yielded lower EMG activity compared to neutral expressions, the same observation was made for fearful facial expressions 3600ms after stimulus onset. Lastly, we did not find a replicable effect of sad facial expressions on the EMG signal (see Fig. 6a below and Table 13 in Online Resource 3). The analyses on the other expression modalities revealed that neither any of the emotional bodily expression nor any of the emotional facial cues had a consistent effect on the Corrugator signal in the training and the test sample.Zygomaticus major. EMG activity over the zygomaticus major region was consistently elevated for happy versus neutral facial expressions starting 700ms after stimulus onset and almost throughout the entire stimulus presentation (700ms %u2013 2600ms, 2800 - 2900ms, 3200 - 3900ms; all ps < .05). Moreover, seeing a fearful facial expressions was related to an enhanced EMG signal 1800-2200ms after stimulus onset in both training and test sample. Activations during the