| | Is a neutral face really evaluated as being emotionally neutral?Received 29 November 2005; received in revised form 26 September 2006; accepted 3 February 2007. Abstract Most of the functional neuroimaging studies on emotion have used neutral faces as a baseline condition. The aim of the present study was to explore whether prototypical neutral faces are evaluated as displaying neutral emotions. Twenty-one subjects performed the Extrinsic Affective Simon Task (EAST), a validated implicit task that measures the emotional evaluation of target stimuli. All stimuli consisted of two juxtaposed faces from standardized facial pictures. The attribute stimuli (positive vs. negative), which needed to be classified on the basis of extrinsic valence, were presented as black and white facial pictures. The target stimuli were color-filtered positive, negative, neutral, and positive/negative faces, and subjects were instructed to classify them on the basis of the filtered color (blue vs. green). The responses to the positive target faces were associated with the positive emotions and the responses to the negative target faces were associated with the negative emotions. For the neutral faces, the responses were similar to those of negative faces, while for the positive/negative stimuli, the responses were undifferentiated. These findings suggested that prototypical “neutral” faces may be evaluated as negative in some circumstances, which suggests that the inclusion of neutral faces as a baseline condition might introduce an experimental confound in functional neuroimaging studies. 1. Introduction  The ability to read emotions from the facial expressions of others is crucial to survival for humans (Darwin, 1872/1965). Some researchers have insisted that emotion is, at least in part, read into the face by the observer (Bruner and Tagiuri, 1954, Russell and Fehr, 1987). Certain configurations of facial features which result from specific patterns of facial muscle contractions have been reported to be universally evaluated as certain basic emotions (Ekman, 1992, Ekman, 1999). These basic emotions include happiness, which is a positive emotion, and sadness, anger, fear, and disgust, which are negative emotions, as well as other emotions such as surprise (Ekman, 1992, Ekman, 1999). To date, most functional neuroimaging studies have conventionally used prototypical neutral faces as a baseline condition for comparing specific facial expressions (Sprengelmeyer et al., 1998, Kesler-West et al., 2001, Pessoa et al., 2002, Kilts et al., 2003; for review, see Phan et al., 2002, Wager and Smith, 2003). However, to our knowledge, no one has addressed whether the prototypical neutral faces are judged to be neutral by subjects during the emotional task of the neuroimaging studies. In most of these studies, the duration and speed of presentation of the emotional stimuli are brief (< 3 s) and fast (the usual interstimulus interval was 2–3 s). In addition, the tasks in numerous studies (Morris et al., 1998, Phillips et al., 1997, Phillips et al., 2001, Phillips et al., 2004) are emotionally implicit, as is indicated by the fact that the limbic regions of the brain are more activated when the task is emotionally implicit (such as gender discrimination) rather than explicit (Phan et al., 2002). In fact, one previous report (Kesler-West et al., 2001) revealed that post-scan ratings of the neutral faces were labeled or identified as happy, angry and fearful, as well as neutral. Although prototypical neutral faces are judged to be neutral because the poser's facial muscles are relaxed (Ekman and Friesen, 1978, Young et al., 1997), the emotional recognition of prototypical neutral faces may be influenced by several factors such as the experimental context (Russell and Fehr, 1987, Surakka et al., 1999). Neutral faces tend to be evaluated as sad if the preceding expression was a happy expression and vice versa (Russell and Fehr, 1987). Self-report may be another influencing factor that is affected by a variety of distortions such as social norms or investigators’ expectations (Parrot and Hertel, 1999). To avoid these pitfalls, implicit measures have been used including the priming method and Greenwald's Implicit Association Tests (IATs) (Fazio and Olson, 2003). Standard IATs assess the strength of associations between a bipolar target concept (flower vs. insect) and a bipolar attribute concept (good vs. bad) by comparing the response times for two differently combined discrimination tasks (Greenwald et al., 1998). The effects produced by IATs are typically much larger, and the reliability is much higher than those reported for priming methods (Egloff and Schmukle, 2002). However, IAT measures are limited to the assessment of the relative strength of associations between bipolar concepts (De Houwer, 2003, Teige et al., 2004). Unlike the IAT, the EAST (Extrinsic Affective Simon Task) (De Houwer, 2003) is a non-relative measure of automatic associations between two concepts, which makes it possible to measure the evaluation of multiple items (name of the participant, first name of previous participant, flower, insect, vs. XXXX) at the same time. The EAST is a variation of the Affective Simon Tasks (De Houwer and Eelen, 1998). In the Affective Simon paradigms, participants are asked to give a positive or a negative response to target stimuli on the basis of a specific stimulus feature (e.g. color) while ignoring the valence of the target stimulus. In an EAST experiment, the attribute stimuli (positive vs. negative) are presented as white words, and target stimuli (e.g. flowers vs. insects) are presented as colored words. Each stimulus appears one-by-one on a monitor. Depending on the word's color (blue vs. green), subjects are instructed to discriminate stimuli by pressing a left or a right response key according to different features. The basic assumption of the EAST is that subjects automatically associate the target stimuli (e.g. flowers) with attribute stimuli (e.g. positive), although they are instructed to concentrate exclusively on color for discrimination. Similar to IATs (usually Cohen's d>0.6), the EAST has a reliable effect size (de Houwer, 2003). Based on the above literature, we realized that it should be possible to investigate the emotional evaluations of neutral expressions by using the pictorial EAST. The pictorial EAST (as well as the word EAST) was expected to provide valid and direct information for effective evaluation (Huijding and de Jong, 2005). The aim of the present study was to determine whether prototypical neutral faces were evaluated as neutral. We measured the emotional evaluation performances of the stimuli of positive, negative and neutral faces with the EAST. Positive/negative stimuli with simultaneous positive and negative valences were added to verify whether the EAST reflects a separate estimate of the emotional evaluation of neutral as well as positive and negative stimuli, since positive and negative faces of positive/negative stimuli will compete and leave relative differences in positive and negative responses undifferentiated. We expected that the responses to the positive and negative faces would be associated with the positive and negative, respectively. We selected happy, neutral, and fearful faces from standardized facial pictures (Ekman and Friesen, 1976) and constructed the positive, negative, neutral, and positive/negative stimuli, which were prepared as juxtaposed happy–happy, fearful–fearful, neutral–neutral, and happy–fearful faces, in an EAST procedure. We hypothesized that this implicit approach would determine the subjective emotional valence of neutral faces better than a traditional approach that could be influenced by the subjects’ intentions or controls (Parrot and Hertel, 1999). 2. Methods 2.1. Subjects Twenty-one healthy subjects (male 9, female 12) were recruited through a local newspaper advertisement. All subjects fulfilled the inclusion criteria, which consisted of their being aged 18 to 40, having more than 12 years of education, and being right-handed. Handedness was assessed with the revised version of Annett's Hand Preference Questionnaire (Annett, 1970, Briggs and Nebes, 1975). All subjects were subjected to a comprehensive psychiatric evaluation using the Structured Clinical Interview for DSM-IV (First et al., 1996). Subjects with additional Axis I or II disorders, neurological disorders, past history of loss of consciousness, or relevant visual defects were excluded. This investigation was carried out in accordance with the Declaration of Helsinki. The study was reviewed and approved by the Institutional Review Board of Severance Mental Health Hospital. All subjects provided written informed consent. The mean age was 25.3 (S.D.=5.1) years, and the mean education was 13.7 (S.D.=1.6) years. Five subjects were married and 16 subjects were unmarried. 2.2. Stimuli Four positive and four negative stimuli were presented in the black and white trials, and another six positive, six negative, six neutral and six positive/negative emotional stimuli were presented in the color-filtered trials. We selected happy, fearful, and neutral faces from the same model (for black and white trials, two male, PE, WF; two female, MF, SW; for the color-filtered trials, three male, EM, GS, JJ and three female, C, MO, NR) from a standard set of prototypical facial expressions (Ekman and Friesen, 1976) using the criterion of well-recognized expressions of happy and fearful emotions from the same number of male and female posers. The juxtaposed happy–happy and fearful–fearful faces were used as positive and negative black and white stimuli. The juxtaposed happy–happy, fearful–fearful, neutral–neutral, and happy–fearful faces were used as color-filtered positive, negative, neutral, and positive/negative stimuli, respectively. The positions of the two faces of each stimulus were counterbalanced. In the color-filtered trials, the faces had a blue or a green colored filter placed over them. The green filtering was created by setting the red, green, and blue values at 0, 25 and 0, and the blue filtering was produced by setting these same values at 0, 0 and 25 in the MGI Photo Suite III program. Each stimulus was presented in random order in all blocks with the restrictions that the same faces should not appear in two or more consecutive trials and that the response keys should not be the same in four or more consecutive trials. Following the proposals of a previous EAST study (De Houwer, 2003, Teige et al., 2004, Huijding and de Jong, 2005), which found that using a more homogeneous set may increase the reliability of the EAST, the trial orders were constant across subjects. The inter-trial interval was 1500 ms. 2.3. Procedures The procedure used in this study was essentially the same as that used in De Houwer's EAST study (De Houwer, 2003), which demonstrates the effectiveness of EAST in assessing a person's emotional evaluation of target stimuli. Subjects were seated in front of a computer monitor with an eye-to-monitor distance of approximately 40 cm. They were asked to classify the facial stimuli by pressing the positive (p) or negative (q) key. In the trials using black and white pictures as attribute stimuli, subjects were requested to press the positive (p) key for the black and white positive faces and to press the negative (q) key for the black and white negative faces. In the trials that used color-filtered pictures as target stimuli, however, they were asked to press the positive or negative key on the basis of the filter color (Table 1). Half of the subjects were told to press the positive key (p) for the blue-filtered pictures and to press the negative key (q) for the green-filtered pictures, while the other half were given a reversed color assignment. After this, the subjects were informed that if they performed an error, a white ‘x’ would appear below the stimuli until they made the correct response. The subjects were asked to respond as accurately, but also as quickly, as possible. Finally, subjects were informed that they were to perform two practice blocks of 48 trials (one black and white block, one color block) and four test blocks of 72 trials (24 black and white trials, 48 color trials). A psychiatrist (KJI) was positioned about 1 m behind subjects and supervised the entire experimental procedure. After the EAST session, all subjects were asked to rate the valence they experienced at the presentation of each facial expression (six happy, six fearful, six neutral faces for color-filtered stimuli and four happy and four fearful faces for black and white trials) using the Self-Assessment Manikin (SAM: Bradley and Lang, 1994), a nine-point scale ranging from 1 (negative) to 9 (positive). 2.4. Data reduction and statistical analysis We analyzed the results of the test trials in which the colored stimuli were presented by only taking into account the accuracy of the initial response as well as the reaction time. According to the routine measurement strategy of the IAT (Greenwald et al., 1998), the reaction time was defined as the elapsed time from the stimulus onset until the correct response was given. Following De Houwer's (2003) procedure, the response latencies below 300 ms were recorded as 300 ms, and latencies above 3000 ms were recorded as 3000 ms. All analyses were based on log-transformed latencies to correct for the skewed latency distribution. Then, the percentage of errors and the mean log-transformed reaction times were calculated separately according to the intrinsic valence and the extrinsic valence (assigned valence of color) of 1) the color stimuli for color-filtered trials in which a positive face was presented and an extrinsically positive response was required (i.e., the response that was assigned to positive black and white stimuli), 2) trials with a positive face and an extrinsically negative response (i.e., the response that was assigned to negative black and white stimuli), 3) trials with negative faces and an extrinsically positive response, 4) trials with negative faces and an extrinsically negative response, 5) trials with neutral faces and an extrinsically positive response, 6) trials with neutral faces and an extrinsically negative response, 7) trials with positive/negative faces and an extrinsically positive response, and 8) trials with positive/negative faces and an extrinsically negative response. A repeated measures analysis of variance (ANOVA) with intrinsic valence (positive, negative, neutral, and positive/negative) and extrinsic valence (positive and negative) of color-filtered stimuli as repeated measures was carried out on percentages of error and reaction times. Wilcoxon signed rank tests were performed on the percentage of error, and post-hoc paired t-tests were performed on the log-transformed reaction times within each intrinsic valence of colored pictures (Fig. 1). The effect size (Cohen's d) of positive (negative) emotion was calculated by the difference of the mean error rates on positive (negative) response trials from that on the negative (positive) response trials divided by the pooled standard deviation of each color-filtered trial. A positive EAST effect size thus signifies that the stimuli were evaluated as positive emotions. The opposite is true for negative emotions. To verify the appropriateness of the facial stimuli, paired t-tests for the ratings of valence of all faces were performed. The criterion for significance was set to P<0.05, and to indicate a trend, P<0.10. 3. Results 3.2. Emotional task (EAST) performance Repeated measures ANOVA, with intrinsic valence (positive, negative, neutral, and positive/negative) and extrinsic valence (assigned color response; positive, negative) of color stimuli as repeated measures, was carried out on error rates and reaction times. Results for error rates indicated no significant main effect of extrinsic or intrinsic valence (For extrinsic valence, F=0.533, d.f.=1,80, Greenhouse–Geisser epsilon=0.007, P=0.467; for intrinsic valence, F=1.847, d.f.=3, Greenhouse–Geisser epsilon=0.065, P=0.145). The interaction effect was significant for intrinsic×extrinsic valence (F=12.590, d.f.=3,80, Greenhouse–Geisser epsilon=0.321, P<0.001). In the Wilcoxon signed rank test (Table 2) in trials of positive target stimuli, subjects performed significantly more accurately when the correct response key was also used to indicate positive valence for black and white pictures than when the response key was assigned to negative black and white stimuli (Z=3.34, P=0.001). In trials of negative stimuli, the reverse was also observed (Z=2.74, P=0.006). In the trials using neutral faces, performance was significantly more accurate when the correct response key was the response key assigned to negative black and white trials (Z=2.68, P=0.007), while in the trials of positive/negative faces, there was no significant difference of performance between the assigned correct response key to positive and negative black and white trials (Z=0.68, P=0.49). The effect size of the EAST signified that the positive, negative, neutral, and positive/negative stimuli were evaluated as positive (Cohen's d=1.98), negative (Cohen's d=−1.14), negative (Cohen's d=−1.03), and undifferentiated (Cohen's d=0.25), respectively. | | |  | Extrinsic valence | Analysis | |
|---|
| Intrinsic valence | Positive | Negative | t/Z | P | |
|---|
| Positive | | | Error rate (%) | 0.79±1.68 | 5.46±4.39 | 3.34 | 0.001 | | | Reaction time (ms) | 756.1±151.6 | 853.4±159.6 | 3.36 | 0.003 | | | Negative | | | Error rate (%) | 5.46±6.34 | 1.59±2.46 | 2.74 | 0.006 | | | Reaction time (ms) | 863±200.9 | 783.9±142.6 | 2.14 | 0.045 | | | Neutral | | | Error rate (%) | 3.69±5.35 | 0.79±1.68 | 2.68 | 0.007 | | | Reaction time (ms) | 818.2±184.7 | 773.7±131.7 | 1.92 | 0.070 | | | Positive/negative | | | Error rate (%) | 1.39±2.41 | 1.89±3.22 | 0.68 | 0.494 | | | Reaction time (ms) | 862.3±221.2 | 809.0±166.1 | 1.33 | 0.198 | | | | |
The error rate difference between positive and negative responses of the neutral stimuli (2.88±5.04) was significantly different from that of the positive/negative stimuli (0.10±5.7) (Z=−2.44, P=0.015). There was no significant effect of the intrinsic valence of the preceding stimuli (F=0.795, d.f.=2.47, Greenhouse–Geisser epsilon=60.721, P=0.481). In addition, the interaction effect of the preceding stimuli valence and the extrinsic response was also not significant (F=6.759, d.f.=2.70, Greenhouse–Geisser epsilon=18.184, P=0.847). The results of repeated measures ANOVA for the reaction time were similar to those of the error rates. The interaction effect was significant (F=5.832, d.f.=3,80, Greenhouse–Geisser epsilon=0.179, P=0.001), but the main effect of extrinsic or intrinsic valence was not significant (for extrinsic valence, F=1.582, d.f.=1,80, Greenhouse–Geisser epsilon=0.019, P=0.212; for intrinsic valence, F=0.254, d.f.=3, Greenhouse–Geisser epsilon=0.009, P=0.858). The results of paired t-tests (Table 2) revealed that in trials with positive stimuli, the subjects performed significantly faster when the correct response key was the response key that was also assigned to positive black and white stimuli than when the response key was assigned to negative stimuli (t=3.36, P=0.003). The reverse was also true in trials of negative stimuli (t=2.14, P=0.045). In the trials of neutral faces, performance was faster at the trend level when the correct response key was also used to indicate the negative valence for black and white trials (t=1.92, P=0.070), while in the trials of positive/negative faces, there was no difference of performance between the positive and negative assigned responses to black and white trials (t=1.33, P=0.198). In addition, the reaction time difference between positive and negative response of the neutral stimuli (44.5±116.9 ms) was not significantly different from that of the positive/negative stimuli (53.3±190.9 ms) (t=−0.27, P=0.79). 3.3. Reliability of the EASTs Following De Houwer (2003), we calculated the split-half reliabilities. The error rates and reaction times for positive, negative, neutral, and positive/negative faces were z-transformed and subsequently averaged separately for the first and second halves of the experiment. The reliabilities of negative and positive/negative faces in reaction time data (Cronbach's α=0.88, 0.74, respectively) were satisfactory while the others were generally low (in reaction time data, Cronbach's α=0.33, 0.20 for positive and neutral faces; in error rates data, Cronbach's α=−0.34, 0.25, 0.24, 0.34 for positive, negative, neutral and positive/negative faces, respectively). 4. Discussion To determine whether the prototypical neutral faces are actually evaluated as neutral, we measured emotional evaluation performances of the positive, negative, neutral, and positive/negative facial stimuli using the EAST, an implicit measure. Upon explicit rating, as expected, the participants judged the positive, negative, and neutral faces as positive, negative, and neutral. In the EAST, positive (happy–happy) and negative (fearful–fearful) stimuli were evaluated as positive and negative, respectively. The EAST responses to the neutral (neutral–neutral) faces resembled those of negative faces, while responses to the positive/negative (happy–fearful) faces were undifferentiated. 4.1. EAST effect of positive and negative faces For positive and negative faces, the subjects’ responses were quicker and more accurate when the correct response key was also used to indicate positive valence for black and white pictures than when the response key was assigned to negative black and white stimuli. In addition, the effect size of the EAST was robust in error rate (for positive, Cohen's d=1.98; for negative, Cohen's d=−1.14), according to Cohen's criteria (Cohen, 1988). These findings were grossly compatible with the previous EAST report using emotional words (De Houwer, 2002, De Houwer, 2003) and emotional pictures (An et al., 2006, Huijding and de Jong, 2005), which showed that the EAST could measure the emotional evaluation of each target emotional stimuli separately and still retain the robust effect size. The EAST procedure with emotional faces in this study may be useful in that positive and negative faces were evaluated as positive and negative, respectively, and also because the effects were robust. 4.2. EAST effect of neutral and positive/negative faces While the performances for the positive/negative stimuli were found to be undifferentiated, the performance for neutral faces mimicked that observed with negative faces, i.e., subjects was significantly more accurate and faster (at a trend level) when the correct response key was associated with negative as compared with positive valence in the black-and-white picture task. In light of the fact that the differences in reaction time, generally the more often cited measure for implicit associations (Fazio and Olson, 2003), our reaction time data for neutral stimuli are moving in the right direction but are only significant at a trend level. However, in a recent EAST study using pictorial stimuli not word stimuli, the error rates were found to be robust (De Houwer, 2003) or more prominent than reaction time (An et al., 2006, Huijding and de Jong, 2005). To provide further evidence supporting the main finding that neutral faces were associated with negative emotion, we directly compared the difference between the positive and negative responses of the neutral trial and the positive/negative trials. With respect to error rates, the difference remained significant, although the reaction time difference was not. The effect size of the EAST on the error rate for neutral faces (Cohen's d=−1.03) was comparable to that for negative faces (Cohen's d=−1.14). It is noteworthy that the responses to neutral faces, which had been validated by prior self-report investigations (Ekman and Friesen, 1976) and also found to be evaluated as a neutral emotion on the SAM rating in the present study, were similar to those to negative faces on EAST performance. There are several possible explanations for this main finding. First, if the EAST effect had a tendency to shift in a negative direction to any kind of stimuli, this tendency might result in evaluation of the neutral faces to be negative. However, a previous study revealed that the EAST yielded a near zero EAST effect to neutral letter strings (i.e., “XXXXX”) (De Houwer, 2003). In addition, the positive/negative faces were found to be undifferentiated in the present study. Therefore, this possible tendency could not explain our main findings. Secondly, the neutral faces might have been perceived as relatively negative in the emotional context of the present EAST. The valence of the preceding emotional stimuli can influence the judgment of valence of the following emotional stimuli (Russell and Fehr, 1987). To control the context effect of the extrinsic and intrinsic valence, the order of the target and attribute stimuli was randomized with the restriction that the same stimuli were not presented on two or more consecutive trials and that the required response was not the same on four or more consecutive trials. Furthermore, we examined the EAST effect on error rates of neutral stimuli according to the intrinsic valence of the preceding stimuli. There was no significant effect of the intrinsic valence of the preceding stimuli and, moreover, the interaction effect of the intrinsic valence of the preceding stimuli and the extrinsic response was also not significant. Therefore, the context effect of the preceding stimuli is unlikely to be an explanation for our main findings. A more plausible explanation is that the neutral faces may really be negatively evaluated, at least in some particular circumstances. The neutral faces in a standard set of facial expressions (Ekman and Friesen, 1976) were validated by identifying or rating the objective emotions of these facial expressions by college students. The neutral faces were identified as objectively neutral, but they might have been perceived as subjectively negative. There was one study showing that emotion judgments of neutral faces were most likely to be negative in nature (Jaeger et al., 1986). They suggested facial muscle movement characterizing the neutral face appears to be similar to that of ‘sad’ emotion. In fact, the prototypical neutral faces (muscles relaxed) can be suggested to be slightly cold and threatening because it is common to expect approval in normal social situations (Phillips et al., 1997, Phillips et al., 2004). One recent functional neuroimaging study (Somerville et al., 2004) suggested that the inherent uncertainty of neutral faces might interact with state anxiety and induce a ventral amygdalar response, which is well known as a fear-related area. In the EAST, emotional stimuli are more briefly and rapidly presented than those of a previous study (Ekman and Friesen, 1976). The response tendency of neutral faces to be subjectively cold and threatening may be exaggerated and more evident when the stimuli were relatively brief and rapidly presented, and were manifested by bypassing social norms and the investigators’ expectations by using implicit measures, as is the case with the EAST. 4.3. EAST reliabilities The reliability of the EAST effects in reaction time data for negative and positive/negative faces were satisfactory (Cronbach's α>0.7), which indicates that the reliabilities might be improved by keeping the presentation order constant across subjects. However, the reliabilities of the EAST effect on error rate data were generally low. These findings were in agreement with previous EAST studies (De Houwer, 2003, Huijding and de Jong, 2005, Teige et al., 2004). The low reliabilities may be due to the assessment of the emotional values of multiple faces simultaneously (Huijding and de Jong, 2005). The low reliabilities indicate that there was a high proportion of erroneous variance in emotional evaluation in this study condition. A high proportion of erroneous variance may lead to an enhanced probability of a type II error and thus may not reflect a true EAST effect. The type I error may not have been affected by low reliability. Instead of a highly reliable emotional task, the EAST task should be considered as an experimental procedure that can assess the emotional values of multiple faces implicitly at the same time. However, on the point that the reliabilities of error rates of EAST performances for both neutral and positive/negative trials were comparable, our main finding that a neutral face was evaluated as a negative emotion in EAST while positive/negative faces were evaluated as undifferentiated might be not entirely derived from the reliability difference between these two neutral and positive/negative trials. 4.4. Limitations There were some important methodological limitations in our study. First, the possibility that the internal state of these subjects had an influence on the evaluation of the neutral faces could not be excluded because the internal states of the perceiver have been found to influence the emotional evaluation, particularly for expressions that are ambiguous or neutral (Innes-Ker and Niedenthal, 2002). We excluded the clinically depressed subjects by using the Structured Clinical Interview for DSM-IV, but we did not examine the internal state of the subjects on the day of the experiment. Secondly, a previous report has shown that the attractiveness of the poser for the stimuli can influence expression–perception of subjects (Höschel and Irle, 2001). We did not investigate the perceived attractiveness of the stimuli. Thirdly, the facial stimuli in this study were all Caucasian faces, which are not as familiar to Korean subjects as Asian faces. There may be cultural and ethnic differences in perceiving facial expressions (Cole et al., 2002) even though basic emotions are reported to be universal (Ekman, 1992, Ekman, 1999). Additionally, there has been a suggestion of in-group advantage in judging emotion across cultures (Elfenbein and Ambady, 2002) and a report that Japanese subjects rated out-group emotion as more negative (Matsumoto, 1990). However, doubts have been raised about methodology to test a possible in-group advantage effect (Matsumoto, 2002). Furthermore, happy, fearful, and neutral faces were verified to be within positive, negative, and neutral ranges according to post-EAST explicit valence rating in our study. Those are the reasons that cultural and ethnic differences could not fully account for the negative evaluation of neutral stimuli in the EAST, although this possibility was not fully excluded. Future studies using Korean facial emotions will be needed to exclude the possible confounding effects of ethnic differences. Lastly, another limitation is that we did not verify the valence of the juxtaposed happy–fearful faces to be near the midpoint on the bipolar SAM rating. However, the distance of the valence of fearful faces (2.44±1.08) from the midpoint (5) was comparable to that of the happy faces (2.65±1.13). The main findings suggest that prototypical “neutral” faces might not be evaluated as neutral. When neutral faces were briefly and rapidly presented, such as in functional neuroimaging studies, they might have been identified negatively, which is a point to consider in the design of future functional neuroimaging studies. According to this implication, it is a plausible alternative to use mildly happy faces derived from morphing 25% happy and 75% neutral faces as the baseline, and this strategy has often been used (Phillips et al., 1997, Phillips et al., 2004). Future studies are required to examine whether mildly happy faces are actually perceived as neutral and also to assess neural responses of emotional stimuli using cross-hair, neutral, and mildly happy faces as control stimuli. Acknowledgement This study was supported by a grant of the Korea Science & Engineering Foundation, interdisciplinary research (Contract grant number: R01–2005–000–10963–0). References An et al., 2006. 1.An SK, Lee E, Kim J-J, Namkoong K, Kang JI, Jeon JH, et al. More impairment on the negative emotion evaluation in patients with paranoid schizophrenia. Yonsei Medical Journal. 2006;47(3):343–353. MEDLINE |
CrossRef
Annett, 1970. 2.Annett M. A classification of hand preference by association analysis. British Journal of Psychology. 1970;61:303–321. MEDLINE Bradley and Lang, 1994. 3.Bradley MM, Lang PJ. Measuring emotion: the self-assessment manikin and the semantic differential. Journal of Behavior Therapy and Experimental Psychiatry. 1994;25:49–59. MEDLINE |
CrossRef
Briggs and Nebes, 1975. 4.Briggs GG, Nebes RD. Patterns of hand preference in a student population. Cortex. 1975;11:230–238. MEDLINE Bruner and Tagiuri, 1954. 5.Bruner JS, Tagiuri R. The Perception of People.. In: Handbook of Social Psychology. vol. 2:Cambridge: Addison-Wesley; 1954;p. 634–654. Cohen, 1988. 6.Cohen J. Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Lawrence Erlbaum; 1988;. Cole et al., 2002. 7.Cole PM, Bruschi CJ, Tamang BL. Cultural differences in children's emotional reaction to difficult situations. Child Development. 2002;73:983–996. MEDLINE |
CrossRef
Darwin, 1872/1965. 8.Darwin C. The Expression of Man and Animals. Chicago: University of Chicago Press; 1872/1965;. De Houwer, 2002. 9.De Houwer J. The implicit association test as a tool for studying dysfunctional associations in psychopathology: strength and limitations. Journal of Behavior Therapy and Experimental Psychiatry. 2002;33:115–133. MEDLINE |
CrossRef
De Houwer, 2003. 10.De Houwer J. The extrinsic affective Simon task. Experimental Psychology. 2003;50:77–85. De Houwer and Eelen, 1998. 11.De Houwer J, Eelen P. An affective variant of the Simon paradigm. Cognition and Emotion. 1998;8:1–20. Egloff and Schmukle, 2002. 12.Egloff B, Schmukle SC. Predictive validity of an implicit association test for assessing anxiety. Journal of Personality and Social Psychology,. 2002;83:1441–1455. MEDLINE |
CrossRef
Ekman, 1992. 13.Ekman P. An argument for basic emotions. Cognition and Emotion. 1992;6:169–200. Ekman, 1999. 14.Ekman P. Facial expressions. In: Dalgleish T, Power M editor. The Handbook of Cognition and Emotion. Sussex: John Wiley and Sons, Ltd.; 1999;p. 301–320. Ekman and Friesen, 1976. 15.Ekman P, Friesen WV. Pictures of Facial Affect. Palo Alto, CA: Consulting Psychologists Press; 1976;. Ekman and Friesen, 1978. 16.Ekman P, Friesen WV. Facial Action Coding System: Investigator's Guide. Palo Alto, CA: Consulting Psychologists Press; 1978;. Elfenbein and Ambady, 2002. 17.Elfenbein HA, Ambady N. On the universality and cultural specificity of emotion recognition: a meta-analysis. Psychological Bulletin. 2002;128:203–235. MEDLINE |
CrossRef
Fazio and Olson, 2003. 18.Fazio RH, Olson MA. Implicit measures in social cognition research: their meaning and use. Annual Review of Psychology. 2003;54:297–327. MEDLINE |
CrossRef
First et al., 1996. 19.First MB, Gibbon M, Spitzer RL, Williams JBW. Structured Clinical Interview for DSM-IV Axis I Disorders: Non-Patients Edition (SCID-I/PS), Version 2. New York: New York State Psychiatric Institute; 1996;. Greenwald et al., 1998. 20.Greenwald AG, McGhee DE, Schwartz JLK. Measuring individual differences in implicit cognition: the Implicit Association Test. Journal of Personality and Social Psychology. 1998;74:1464–1480. MEDLINE |
CrossRef
Höschel and Irle, 2001. 21.Höschel K, Irle E. Emotional priming of facial affect identification in schizophrenia. Schizophrenia Bulletin. 2001;27:317–327. MEDLINE Huijding and de Jong, 2005. 22.Huijding J, de Jong PJ. A pictorial version of the extrinsic affective Simon tasks: sensitivity to generally affective and phobia-relevant stimuli in high and low spider fearful individuals. Experimental Psychology. 2005;52:289–295. Innes-Ker and Niedenthal, 2002. 23.Innes-Ker A, Niedenthal PM. Emotion concepts and emotional states in social judgment and categorization. Journal of Personality and Social Psychology. 2002;83:804–816. MEDLINE |
CrossRef
Jaeger et al., 1986. 24.Jaeger J, Borod JC, Peselow E. Facial expression of positive and negative emotions in patients with unipolar depression. Journal of Affective Disorders. 1986;11:43–50. Abstract | Full Text |
Full-Text PDF (252 KB)
|
CrossRef
Kesler-West et al., 2001. 25.Kesler-West ML, Andersen AH, Smith CD, Avison MJ, Davis CE, Kryscio RJ, et al. Neural substrate of facial emotion processing using fMRI. Cognitive Brain Research. 2001;11:213–226. MEDLINE |
CrossRef
Kilts et al., 2003. 26.Kilts CD, Egan G, Gideon DA, Ely TD, Hoffman JM. Dissociable neural pathways are involved in the recognition of emotion in static and dynamic facial expressions. NeuroImage. 2003;18:156–168. MEDLINE |
CrossRef
Matsumoto, 1990. 27.Matsumoto D. Cultural similarities and differences in display rules. Motivation and Emotion. 1990;13:195–214. Matsumoto, 2002. 28.Matsumoto D. Methodological requirements to test a possible in-group advantage in judging emotions across cultures: comment on Elfenbein and Ambady (2002) and evidence. Psychological Bulletin. 2002;128:236–242. MEDLINE |
CrossRef
Morris et al., 1998. 29.Morris JS, Ohman A, Dolan RJ. Conscious and unconscious emotional learning in the human amygdala. Nature. 1998;393:467–470. MEDLINE |
CrossRef
Parrot and Hertel, 1999. 30.Parrot WG, Hertel P. Research methods in cognition and emotion. In: Dalgleish T, Power M editor. The Handbook of Cognition and Emotion. Sussex: John Wiley and Sons, Ltd.; 1999;p. 61–82. Pessoa et al., 2002. 31.Pessoa L, McKenna M, Gutierrez E, Ungerleider LG. Neural processing of emotional faces requires attention. Proceedings of the National Academy of Sciences of the United States of America. 2002;99:11458–11463. MEDLINE |
CrossRef
Phan et al., 2002. 32.Phan KL, Wager T, Talyor SF, Liberzon I. Functional neuroanatomy of emotion: a meta-analysis of emotion activation studies in PET and fMRI. Neuroimage. 2002;16:331–348. MEDLINE |
CrossRef
Phillips et al., 1997. 33.Phillips ML, Young AW, Senior C, Brammer M, Andrew C, Calder AJ, et al. A specific neural substrate for perceiving facial expressions of disgust. Nature. 1997;389:495–498. MEDLINE |
CrossRef
Phillips et al., 2001. 34.Phillips ML, Medford N, Young AW, Williams L, Williams SCR, Bullmore ET, et al. Time course of left and right amygdalar responses to fearful facial expressions. Human Brain Mapping. 2001;12:193–202. MEDLINE |
CrossRef
Phillips et al., 2004. 35.Phillips ML, Williams LM, Heining M, Herba CM, Russell T, Andrew C, et al. Differential neural responses to overt and covert presentations of facial expressions of fear and disgust. Neuroimage. 2004;21:1484–1496. MEDLINE |
CrossRef
Russell and Fehr, 1987. 36.Russell JA, Fehr B. Relativity in the perception of emotion in facial expressions. Journal of Experimental Psychology: General. 1987;116:223–237.
CrossRef
Somerville et al., 2004. 37.Somerville LH, Kim H, Johnstone T, Alexander AL, Whalen PJ. Human amygdala responses during presentation of happy and neutral faces: correlations with state anxiety. Biological Psychiatry. 2004;55:897–903. Abstract | Full Text |
Full-Text PDF (269 KB)
|
CrossRef
Sprengelmeyer et al., 1998. 38.Sprengelmeyer R, Rausch M, Eysel UT, Przuntek H. Neural structures associated with recognition of facial expressions of basic emotions. Proceedings. Biological Sciences/The Royal Society. 1998;265:1927–1931. Surakka et al., 1999. 39.Surakka V, Sams M, Hietanen JK. Modulation of neutral face evaluation by laterally presented emotional expressions. Perceptual and Motor Skills. 1999;88:595–606. MEDLINE |
CrossRef
Teige et al., 2004. 40.Teige G, Schnabel K, Banse R, Asendorpf JB. Assessment of multiple implicit self-concept dimensions using the Extrinsic Affective Simon Task (EAST). European Journal of Personality. 2004;18:495–520. Wager and Smith, 2003. 41.Wager TD, Smith EE. Neuroimaging studies of working memory: a meta-analysis. Cognitive, Affective and Behavioral Neuroscience. 2003;3:255–274. MEDLINE |
CrossRef
Young et al., 1997. 42.Young AW, Rowland D, Calder AJ, Etcoff NL, Seth A, Perrett DI. Facial expression megamix: tests of dimensional and category accounts of emotion recognition. Cognition. 1997;63:271–313. MEDLINE |
CrossRef
a Department of Psychiatry, Ilsan Hospital, National Health Insurance Corporation, Gyeonggi-do, South Korea b Department of Psychiatry, Yonsei University College of Medicine, Seoul, South Korea c Institute of Behavioral Science in Medicine, Yonsei University College of Medicine, Seoul, South Korea  Corresponding author. Department of Psychiatry, Severance Mental Health Hospital, Yonsei University College of Medicine, 696–6 Tanbul-dong, Gwangju-si, Gyeonggi-do, 464–100, South Korea. Tel.: +82 31 760 9404; fax: +82 31 761 7582.
PII: S0165-1781(07)00035-2 doi:10.1016/j.psychres.2007.02.005 © 2007 Elsevier Ireland Ltd. All rights reserved. | |
|
FULL TEXT00035-2/journalimage?img=PIIS0165178107000352.gr1.lrg.gif&fig=fig1&kwhquery=&issn=0165-1781&ishighres=false&allhighres=false&free=yes','journalimage');)
