تاثیر شدت بر حساسیت کودکان به حالات چهره شاد، غمگین و ترس

Abstract Most previous studies investigating children’s ability to recognize facial expressions used only intense exemplars. Here we compared the sensitivity of 5-, 7-, and 10-year-olds with that of adults (n = 24 per age group) for less intense expressions of happiness, sadness, and fear. The developmental patterns differed across expressions. For happiness, by 5 years of age, children were as sensitive as adults even to low intensities. For sadness, by 5 years of age, children were as accurate as adults in judging that the face was expressive (i.e., not neutral), but even at 10 years of age, children were more likely to misjudge it as fearful. For fear, children’s thresholds were not adult-like until 10 years of age, and children often confused it with sadness at 5 years of age. For all expressions, including even happy expressions, 5- and 7-year-olds were less accurate than adults in judging which of two expressions was more intense. Together, the results indicate that there is slow development of accurate decoding of subtle facial expressions.

Introduction Facial expressions are an important source of social information. Accurate recognition of facial expressions allows us to make inferences about other people’s feelings, thereby guiding our social behavior. Normal human adults are fast and accurate at recognizing facial expressions, even from still photographs (Ekman, 1993). This ability is seen universally; there is high agreement among adults from different cultures on what emotion is shown in still photographs of facial expressions of basic emotions (happiness, sadness, anger, surprise, fear, and disgust) (Ekman and Friesen, 1971, Ekman et al., 1987, Elfenbein and Ambady, 2002 and Izard, 1971). However, the development of this ability remains largely unclear. Most previous studies investigating children’s ability to recognize facial expressions used photographs of intense facial expressions of basic emotions. Children’s performance in identifying emotion from such photographs improves with age (reviewed in Herba & Phillips, 2004), with positive expressions recognized earlier and more accurately than negative expressions (Boyatzis et al., 1993, Camras and Allison, 1985 and Widen and Russell, 2003). Overall, the improvement can be characterized by a large increment in accuracy between 3 and 7 years of age (Camras and Allison, 1985, Durand et al., 2007, Markham and Wang, 1996 and Vicari et al., 2000) and an increment in speed between 7 and 10 years of age (De Sonneville et al., 2002). The developmental patterns for intense emotional expressions are similar across studies that used photographs of children’s faces (Boyatzis et al., 1993, Camras and Allison, 1985 and Widen and Russell, 2003), photographs of adults’ faces (Durand et al., 2007, Markham and Wang, 1996 and Vicari et al., 2000), or both (De Sonneville et al., 2002). At least with some stimulus sets, there is continued improvement in accuracy into early adolescence (Kolb, Wilson, & Taylor, 1992). Moreover, children’s pattern of brain activation when processing different intense facial expressions differs from that of adults until at least 11 years of age for functional magnetic resonance imaging (fMRI) activation to neutral versus fear (Thomas et al., 2001; see also Monk et al., 2003) and until late adolescence for event-related potential (ERP) patterns (Batty & Taylor, 2006). The previous studies with intense emotional expressions documented that children are accurate in judging intense exemplars by approximately 7 years of age, with subsequent changes in reaction time and neural specificity. However, in everyday life, we see less intense facial expressions more frequently than intense facial expressions. The ability to recognize less intense facial expressions and subtle changes in the intensity of facial expressions (e.g., to see that someone is mildly amused by a joke) facilitates smooth social interactions. Therefore, it is important to investigate children’s ability to recognize facial expressions of lower intensity. The intensity of a facial expression is determined by the amount of muscle displacement away from a neutral state (Hess, Blairy, & Kleck, 1997). For example, the intensity of a happy expression can be characterized by the degree of displacement of zygomaticus major and orbicularis oculi muscles relative to their relaxed states (Duchenne de Boulogne, 1990). Three recent studies of children used a morphing process to move the positions of features in a neutral face toward their positions in an intense emotional face, a change simulating the consequences of facial muscle movements. One study compared children with and without psychopathic tendencies in a program for troubled children (Blair, Colledge, Murray, & Mitchell, 2001). Children with psychopathic tendencies needed significantly more intensity to recognize the sad expression, and they were more likely to mistake the fearful expression for another expression even at full intensity. However, the children spanned the age range of 9 to 17 years, and the authors did not investigate the effect of age on thresholds or errors. In a second study, Herba and colleagues (2008; see also Herba, Landau, Russell, Ecker, & Phillips, 2006, for related results on matching expression across intensity) used 10 levels of intensity to investigate the effect of familiarity on 4- to 15-year-olds’ perception of five facial expressions (happiness, sadness, anger, fear, and disgust) in familiar and unfamiliar adult faces. Sensitivity improved with increasing ages for happy and fearful expressions but not for disgust, sad, and angry expressions, with no facilitation by familiarity for any facial expression and in fact some evidence that familiarity degraded sensitivity. However, because there was no adult comparison group, it is not possible to determine when sensitivity reaches adult levels. A different pattern emerged in a recent study that used morphing to create six intermediate intensities between neutral and expressions of fear and anger; children (7–13 years of age) and adolescents (14–18 years of age) were less sensitive than adults for both anger and fear (Thomas, De Bellis, Graham, & LaBar, 2007). These data suggest that the development of sensitivity to at least some facial expressions continues into adolescence. However, the authors used wide age groupings and did not analyze misidentifications. Adults tend to make systematic confusions among facial expressions. For example, they often confuse fear with surprise and also confuse anger with disgust (Ekman and Friesen, 1971, Etcoff and Magee, 1992 and Young et al., 1997). Little information is known about whether children show the same pattern of confusion among facial expressions as do adults. The purpose of our study was to build on these previous findings by including more intensity levels and a method that allowed us to measure both thresholds for each expression and confusions among expressions. We systematically manipulated the intensity of three facial expressions (happiness, sadness, and fear) by morphing photographs of intense exemplars of these expressions with photographs of neutral faces of the same models to create 20 levels of intensity. With intensity of expression as a factor in the experiment, accuracy is not an adequate measure of children’s performance because participants can make two types of errors. The first type of error is specific to low-intensity expressions. Children, as well as adults, may fail to detect any expression in a face when the intensity of that expression is very low. To measure this type of error, we calculated thresholds to detect expressions in faces, that is, to see that the expression is not neutral. The second type of error, as in studies using only intense facial expressions, is to misidentify one expression as another. To measure this type of error, we calculated the percentage of misidentifications in faces that were recognized as emotional. One concern is whether the values are affected by the particular choices of expression presented to the participants, in this case happiness, sadness, and fear. We suspect that the threshold measure will not be affected by the particular choices because it measures the intensity at which the expression is seen as no longer neutral even if the expression cannot be identified correctly at that low intensity level. However, the specific choices will affect the pattern of misidentifications because the forced-choice procedure limits the types of errors that can occur. In the discussion, we consider the effect of this limitation on the interpretation of our findings. We chose to study happy and sad expressions because, with intense expressions, children show adult-like accuracy for them earlier than they do for other expressions (Boyatzis et al., 1993, Camras and Allison, 1985 and Widen and Russell, 2003), perhaps because of relatively greater exposure to happy and sad faces in everyday life. We chose fear as the third expression because it is likely to show a different developmental trajectory. With intense exemplars, adult-like sensitivity to fear develops relatively late (Camras and Allison, 1985, Durand et al., 2007, Markham and Wang, 1996 and Vicari et al., 2000), possibly because of low exposure in everyday life. From an evolutionary perspective, the late development of adult-like accuracy for fear is surprising because fear signals potential environmental threat and the need to take action to avoid the threat. Consistent with this evolutionary perspective, by 7 months of age infants generalize habituation across different individual faces showing fearful expression (Nelson & Dolgin, 1985) and they look longer at fearful faces than at happy faces. This early onset of processing of fearful faces may be related to the functioning of a specific brain circuit involving the amygdala (Adolphs, Tranel, Damasio, & Damasio, 1995). In human adults, the amygdala can be activated by exposure to fearful faces through a fast and seemingly automatic response to low spatial frequency information carried through subcortical connections as well as by slower cortical input that is likely to develop later (Vuilleumier, Armony, Driver, & Dolan, 2003). Here we investigated whether the developmental trajectories for happiness, sadness, and fear for less intense exemplars are similar to those reported previously for intense expressions. The same method could be used in future studies to explore the development of sensitivity to the other basic emotions (anger, surprise, and disgust). We did not do so here because we were worried that including additional expressions would make the task too complex for the youngest children. In Experiment 1, we investigated age differences in sensitivity to happy, sad, and fearful expressions in 5-, 7-, and 10-year-olds and a comparison group of adults. We also investigated confusions among these expressions at intensity levels above threshold. In Experiment 2, we investigated children’s ability to distinguish between different intensities of the same expression. We expected that, unlike the results of previous studies using only intense expressions, young children would differ from adults in the detection and discrimination of all three facial expressions and would more often misidentify expressions of moderate intensity. For each experiment, we developed a child-friendly procedure suitable for children as young as 5 years of age. Experiment 1 In Experiment 1, we presented photographs of neutral faces and of happy, sad, and fearful faces with 20 levels of intensity to adults and children (7–10 years of age) who were instructed to sort the photographs into four categories: happiness, sadness, fear, and neutrality. In addition, because pilot work indicated that 5-year-olds were not able to respond consistently during such a long procedure, we tested a group of 5-year-olds with 10 of the intensity levels.