سطح انگیختگی حالت چهره تقلید صورت را تعدیل می کند

Abstract We investigated the effect of facial expression arousal level and mode of presentation on facial mimicry. High- and low-arousal facial expressions indicating pleasant and unpleasant emotions were presented both statically and dynamically. Participants' facial electromyographic (EMG) reactions were recorded from the zygomatic major and corrugator supercilii muscles. Stronger zygomatic major muscle activity was evoked for high- compared to low-arousal pleasant expressions. Comparable activity was induced in the corrugator supercilii muscle in response to both high- and low-arousal unpleasant expressions, and this was true for both dynamic and static presentations. These results suggest that the arousal levels of pleasant, but not unpleasant, facial expressions can enhance facial mimicry.

1. Introduction People often produce facial movements congruent with others when looking at emotional facial expressions. This phenomenon is known as facial mimicry and plays an important role in nonverbal communication (Dimberg, 1982). Facial mimicry has been studied in the context of a wide variety of fields, including social factors (McHugo et al., 1991 and Vrana and Gross, 2004), psychotherapy (Rogers, 1957), and emotional contagion (Hatfield et al., 1993). Facial muscle activity evoked in observers is usually recorded using facial electromyography (EMG), which can detect subtle facial movements that are not discernible to the naked eye. Previous studies have demonstrated that photographic presentation of happy facial expressions induces zygomatic major muscle activity (pulling of lip corners, prototypical in happy facial expressions) in observers, whereas angry facial expressions evoke corrugator supercilii muscle activity (lowering of the brows, prototypical in angry facial expressions) (Dimberg, 1982 and Dimberg, 1997). The majority of previous studies focusing on facial mimicry have employed happy and angry facial stimuli as pleasant and unpleasant expressions. These studies are based on the discrete emotion theory (Ekman, 1972) that suggests six basic facial expressions exist. The various theories regarding emotional facial expressions remain disputed (e.g., Scherer and Ellgring, 2007; for a review, see Russell et al., 2003). One theory proposes that facial expressions are interpreted based on two dimensions, their emotional valence (i.e., indicating pleasure or unpleasure) and degree of arousal (Russell and Bullock, 1985). Accordingly, the dimensions of valence and arousal might be important factors for understanding facial mimicry. Happy and angry facial expressions represent high-arousal levels in the dimensional view (Russell and Bullock, 1985), and it has been shown that happy and angry expression stimuli used in several previous studies on facial EMG activity (Ekman and Friesen, 1976) were indeed rated as high arousal (Vrana and Gross, 2004). Therefore, the results of these studies only address facial mimicry with respect to high-arousal emotional expressions. A study has suggested that arousal corresponds to the intensity of emotion (Lang et al., 1998). Given that the high-arousal leads to strong emotional intensity of facial expressions, greater facial EMG activity would be expected to be induced in response to high-arousal than low-arousal facial expressions. Previous studies found comparable facial EMG activity in response to sad and angry facial expressions (Lundqvist and Dimberg, 1995 and Sonnby-Borgström et al., 2008). Sad facial expressions are normally thought to indicate low-arousal levels (Russell and Bullock, 1985), and hence these data suggest that high- and low-arousal unpleasant expressions have similar potential for facial mimicry. However, these experiments did not quantify the arousal level of their stimuli, and thus the possibility exists that some sad face stimuli could indicate high-arousal/intensity levels (e.g., Anttonen et al., 2009 and Harrison et al., 2007). As such, whether high- and low-arousal unpleasant facial stimuli induce comparable facial EMG activity remains unclear. Furthermore, whether facial EMG activity in response to pleasant facial expressions differs between high and low-arousal levels is unknown. Consequently, no extant studies have systematically examined the relationship between arousal level and facial mimicry by employing high- and low-arousal pleasant and unpleasant facial expressions as stimuli. Many studies have suggested that the arousal level of emotional events can modulate facial EMG reactions. For example, Greenwald et al. (1989) recorded facial EMG activity in response to emotional scenes that varied widely across the dimensions of valence and arousal. They reported a trend toward a positive association between arousal ratings and zygomatic major (but not corrugator supercilii) muscle activity. Witvliet and Vrana (1995) assessed facial EMG activity while subjects imagined emotional events with high- and low-arousal levels as well as pleasant and unpleasant meanings. They found that zygomatic major muscle activity was higher during pleasant, high-arousal conditions than pleasant, low-arousal conditions, whereas corrugator supercilii activity was not. Based on these data, we hypothesized that zygomatic major muscle activity is greater for high-arousal than low-arousal pleasant facial expression stimuli. In addition to the static facial expression stimuli used in several previous studies, we also tested dynamic facial expression stimuli. As dynamic facial expressions represent the natural form of communication in daily life, the use of these stimuli increases the ecological validity of our results. Consistent with this notion, recent studies have shown that compared to static expressions, presentation of dynamic facial expressions enhanced facial EMG reactions without any qualitative changes (Sato et al., 2008 and Weyers et al., 2006). Based on these data, we hypothesized that the aforementioned effect of arousal on facial mimicry would be evident and qualitatively comparable for dynamic and static facial expressions. To test these hypotheses, we presented high- and low-arousal facial expressions indicating pleasant and unpleasant emotions in both static and dynamic presentation. For dynamic stimuli, the sequence from a neutral to full-blown expression was presented. For static stimuli, a full-blown expression was presented as a still. EMG activity was recorded from the zygomatic major and corrugator supercilii muscles.

3. Results 3.1. Mean rating of experienced emotions Table 1 shows subjective emotional states when the participants were viewing each facial stimulus. To confirm that the emotion elicited by these facial stimuli was differentiated with respect to valence and arousal, one-tailed t-tests were conducted for each pair of stimuli. Pleasant expressions were rated as significantly more pleasant than unpleasant expressions under both dynamic and static conditions (high-arousal expressions: ts(31) > 15.07, ps < 0.01; low-arousal expressions: ts(31) > 9.88, ps < 0.01). High-arousal expressions were rated as significantly more aroused than low-arousal expressions under both presentation conditions (pleasant expressions: ts(31) > 3.51, ps < 0.01; unpleasant expressions: ts(31) > 8.54, ps < 0.01). The results demonstrate that the facial stimuli expressed the intended emotions to the participants. Table 1. Mean (with SE) rating of experienced emotion (N = 32). Ratings Presentation mode Facial stimuli Pleasantness Unpleasantness High-arousal Low-arousal High-arousal Low-arousal Valence Dynamic 7.59 (0.18) 6.57 (0.21) 2.49 (0.18) 3.23 (0.15) Static 7.55 (0.18) 6.66 (0.20) 2.41 (0.19) 3.42 (0.17) Arousal Dynamic 6.94 (0.17) 4.23 (0.21) 6.57 (0.19) 4.23 (0.15) Static 6.91 (0.18) 4.62 (0.18) 6.63 (0.17) 4.26 (0.21) Note. Scores ranged from 1 (unpleasantness or low arousal) to 9 (pleasant or high arousal). Table options 3.2. Zygomatic major Zygomatic major muscle activity for each expression is shown in Fig. 2(a). The three-way ANOVA showed a significant main effect of valence (F(1, 31) = 5.73, p < 0.05, ηp2 = 0.16), indicating a significantly different pattern of zygomatic major activity for pleasant and unpleasant expressions. Significant interactions were observed between valence and arousal level (F(1, 31) = 4.32, p < 0.05, ηp2 = 0.12) and valence and condition (F(1, 31) = 4.42, p < 0.05, ηp2 = 0.013). Other main effects and interactions were not significant (Fs(1, 31) < 2.60, n.s.). Means EMG change activity and standard errors for (a) zygomatic major muscle and ... Fig. 2. Means EMG change activity and standard errors for (a) zygomatic major muscle and (b) corrugators supercilii. Figure options Follow-up ANOVAs revealed that high-arousal pleasant expressions induced significantly stronger zygomatic major activity than low-arousal pleasant expressions (F(1, 62) = 3.15, p < 0.05), and that pleasant expressions evoked greater zygomatic major activity than unpleasant expressions under the high-arousal condition (F(1, 62) = 13.13, p < 0.01), but not the low-arousal condition (F(1, 62) = 0.19, n.s.). With respect to presentation condition, pleasant expressions produced significantly greater zygomatic major activity compared to unpleasant expressions under the dynamic condition (F(1, 62) = 9.46, p < 0.01), but not the static condition (F(1, 62) = 1.48, n.s.) 3.3. Corrugator supercilii Fig. 2(b) shows corrugator supercilii muscle activity for each expression. The ANOVA yielded the significant main effects of valence (F(1, 31) = 6.69, p < 0.05, ηp2 = 0.18). The interaction between presentation condition and valence was also significant (F(1, 31) = 6.35, p < 0.05, ηp2 = 0.17). Follow-up ANOVAs showed that unpleasant expressions induced significantly greater corrugator muscle activity than pleasant expressions under the dynamic condition (F(1, 62) = 11.10, p < 0.01), but not the static condition (F(1, 62) = 2.18, n.s.). Other main effects and interactions did not reach significance (Fs(1, 31) < 2.09, n.s.)