SEARCH

SEARCH BY CITATION

Keywords:

  • correctness;
  • emotion;
  • facial expression;
  • intensity;
  • recognition

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Using the Japanese and Caucasian Facial Expressions of Emotion (JACFEE) photo set, the relationship between recognition and intensity ratings of universal facial expressions of emotions in 123 Japanese undergraduate students was examined and compared with data reported by American raters. In Japanese raters, although the intensity was rated as high for some of the poses, their correctness scores were poor, suggesting a serious misjudgment of the intended emotions as defined in the JACFEE photo set. Only in Japanese raters were significant relationships between the intensity scores and the percentage correctness scores for sadness detected (r = 0.97, P < 0.0001), but no significant relationship was observed for other emotions. The robust correlation suggests the possibility that Japanese raters might be more responsive to certain emotional expressions when they are fully or intensely expressed. It is proposed that the facial emotional expression paradigm cannot be applied to the psychiatric setting without first refining for cultural differences.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Recently, the study of faces has become popular. These studies range from the investigation of beauty and attraction, 1 to the psychiatrically relevant topics of percentage changes of emotional expressions in depression and schizophrenia. 2 Cross-cultural issues have been addressed extensively, 3–6 while high technological approaches include the identification of brain structures responsible for registering fear or aggression. 7,8

In order to identify the most suitable paradigm for investigating psychopathology and perceptual changes in Japanese psychiatric subjects, we tested the Japanese and Caucasian Facial Expressions of Emotion (JACFEE) in a sample of young Japanese subjects. 6 The JACFEE photo set 9 was developed by American workers and has been used by many investigators in the field of facial research. Some cross-cultural results have been reported 3,4 and these analyses indicated a significant rater culture and emotion factor interaction. Our data showed even more dramatic ethnic differences, 6 which encouraged us to further examine the still photo paradigm for testing emotional expressions. It also alerted us to the fact that much work is needed before one embarks upon studying the emotionally compromised psychiatric patient.

We were interested by the degree of incorrectness in the identification of facial emotional expressions by our Japanese raters under the Matsumoto–Ekman still photo paradigm. Our previous data show that out of the seven emotions posed, the Japanese raters did very poorly (45–56% correct) in recognizing anger, contempt, and fear. They did poorly in recognizing disgust and sadness (64–67% correct), and only scored well for recognition of surprise and happiness (89–95% correct). This poor performance of Japanese normal control raters prompted us to further examine their scores. It has been commented by many that Japanese society discourages overt expression of emotions, particularly negative emotions. The collectivistic Japanese family culture emphasizes the control of anger and other negative emotional expressions in public and sees it as a way of facilitating a harmonious community. In contrast, the individualistic American culture is conducive to the open expression of emotions and places high value on assertiveness. This polarized attitude towards the expression of one’s inner emotional state should also influence the observer’s own emotional reaction when confronted with others’ negative emotional expressions and their perception of others’ emotional expression. For example, the discouragement of overt emotional expression in the Japanese culture is so deeply ingrained that some investigators commented on the possibility of an unconscious denial when a Japanese person is confronted with a face that is expressing negative emotions. This could potentially introduce serious experimental bias when emotionally compromised psychiatric subjects are tested for cognitive changes using the facial emotional expression paradigm.

We report our analysis of the relationship between correctness of recognition and the intensity of rated emotions by our Japanese rater sample. We compared the results with published data on American raters obtained from the literature and the JACFEE photo set. We propose that cultural differences in attitude towards negative emotional expressions could potentially confound the data and such culture-specific reactions to negative emotional reactions should be further investigated before subjects with compromised cognitive ability, such as psychiatric subjects, are tested.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

The study design has been described previously. 6 Briefly, 123 Japanese subjects, consisting of 42 males and 81 females (mean age, 19.8 years), were recruited from the Shiga University of Medical Science, Otsu, Japan. The subjects were shown the JACFEE photo set for 10 s, which comprises 56 photos, including eight photos each of anger, contempt, disgust, fear, happiness, sadness, and surprise. Four photos of each emotion depicted subjects of either Japanese or Caucasian descent (two men, two women). The subjects rated the expressions from a choice of seven emotions and an intensity rating of nine points (0–8) scale: neutral (0), weak (1), moderate (4) and strong (8). The results were published elsewhere. 6

In order to examine the relationship between the intensity and the correctness score, we constructed correlation graphs for each of the seven emotions. There were eight photos for each emotion and the intensity rating scores were plotted against the correctness percentage scores. Scores generated by our Japanese raters were compared with the scores reported by American raters using the JACFEE. 4,5,9 The US-born American sample included 128 men (mean age, 19.9 years) and 143 women (mean age, 19.8 years) and recruited from universities in the San Francisco Bay area. All were at least third generation American and no-one was of Asian descent. 4,5,9

The scores for intensity and percentage correctness were plotted for each emotional expression. Statistical analysis for correlation and significance was performed using the SPSS statistical package (SPSS Inc., Chicago, USA).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

The analysis showed some interesting differences between the Japanese and American raters. With a score of 4 being the median score for intensity and 70% correctness score defined as the significant threshold for universal emotional recognition, 3–5 the ratings generated by these two groups of culturally distinct raters could be compared ( Figs 1,2).

image

Figure 1. . Evaluation of seven emotional expressions from Japanese raters reported by Shioiri et al.6 Abscissa and ordinate indicate the percentage of accuracy (correct percentage) and intensity ratings, respectively. In accordance with Matsumoto, 70% was regarded as the borderline of universal emotion. 5 We also established the intermediate value for intensity and lined this value at 4.

Download figure to PowerPoint

image

Figure 2. . Evaluation of seven emotional expressions from American raters calculated from the literature. 4,5,9

Download figure to PowerPoint

For American raters, with the exception of only one emotion (contempt) and one pose (one of the eight poses for anger), all the scores for the other six emotions and poses of each emotion were tightly segregated, mostly in the north-east segment of the graph. Their correctness scores were not related to their intensity rating scores.

The Japanese raters showed the same phenomenon for two of the seven emotions, namely surprise and happiness. The score for three emotions (anger, fear and disgust), were largely in the north-west segment ( Fig. 1). Although they rated the intensity as high for some of the poses, their correctness scores were poor. This indicated a serious misjudgment of the intended emotions as defined in the JACFEE photo set. An interesting comparison came from two emotions, contempt and sadness. For contempt, while the American raters did not score as well as for the other emotion poses, their scores were tight for seven of the eight poses. The Japanese raters, however, appeared to achieve better percentage correct scores with the poses they identified as intense. The correlation almost reached significance (r = 0.69, P < 0.06). For sadness, while the American raters gave tightly segregated scores, our Japanese raters returned an almost perfect correlation between the percentage correct scores and the intensity scores (r = 0.97, P < 0.0001).

There was a significant difference in sex ratio between Japanese and American subjects (χ2 = 5.91, P < 0.02). In Japanese subjects, however, we found no significant effects of sex to the correct percentage scores and intensity scores in each emotion. As for American subjects, we had no available data to examine the sex difference because we used the results reported by a previous study. 4,5,9

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Matsumoto suggested that universal recognition of emotional expressions should fulfill two criteria: percentage of judges identifying the target emotion is greater than chance and is true across cultures. 5 Our analysis of Japanese subjects ( Fig. 1) demonstrated that in applying these criteria, the emotions of anger, contempt, and fear did not make the threshold for universality. The scores for the emotion disgust barely reached the 70% cut-off threshold. This was irrespective of the perceived intensity of the posed emotions in many cases. Fear, for example, was rated with relatively high intensity but the identification of the emotion was poor. Contempt had both poor recognition and low intensity ratings. The recognition of sadness depended upon the perceived intensity of the pose.

Pilowsky and Katsikitis proposed that emotions differ in their configurational properties. 10 Emotions such as happiness and surprise were postulated to be easier to recognize as these were ‘hardwired’, while the ability to recognize or discriminate between emotions such as anger, disgust and fear requires socialization. Socialization implies cultural influences. If the Japanese raters were exposed less frequently to ‘negative emotions’ in their socialization process, their ability to recognize such emotions could be compromised. Such unfamiliarity with ‘negative emotion’, which in turn has an underlying complexity (i.e., fear has an element of surprise) would be expected to lead to a higher percentage of incorrect scores. This was observed. These data would support the general impression and Japanese family teaching that negative emotional expressions are not encouraged in Japanese society.

Of the seven emotions posed, happiness is the only positive emotion and surprise may be seen as neither positive nor negative. While sadness is not a positive emotional state, it differs from anger, contempt, disgust and fear, in that it is an emotional expression that might draw sympathy from others. As such, it may not elicit a negative response from the observer as compared with the other emotional expressions. In addition, expression of sadness in Japan is more widely accepted compared with the expression of negative emotions. In summary, it appears that the Japanese raters experienced difficulties with emotional expressions that were negative or would elicit a negative reaction. They did not have difficulty with positive or neutral emotional expressions.

It should be noted that some investigators are of the opinion that emotional expressions are not clearly universal. 11 Russell and Bullock found that raters had difficulty restricting some facial expressions to only one category because the taxonomy of emotions was unclear. 12 Nummenmaa discussed the difficulty in recognizing facial expressions in still photographs. 13 In a society such as Japan, where restraints in overt emotional expressions are emphasized, the still photo paradigm may have serious limitations. The robust correlation between the intensity scores and the percentage correctness scores for sadness alerts us to examine the possibility that Japanese raters might be more responsive to certain emotional expressions when they are fully or intensely expressed. This was not true for the American raters.

We should also refer to the limitation of the method. We used the operational definition of ‘correct’ facial expression recognition because definition of absolute ‘correctness’ is impossible. Moreover, the correct percentage of Japanese subjects was very low, except for happiness and surprise. These results might indicate poor validity of the method using the JACFEE in Japanese raters rather than poor capacity of them.

In summary, the facial emotional expression paradigm cannot be applied to the psychiatric setting without first refining it for cultural differences. The still photo approach may meet with difficulty for Japanese raters. It is important to construct photos with higher correctness scores for Japanese raters for experimental purposes. For the emotionally compromised psychiatric patients, blindly presenting facial emotional expression photos without cultural adaptation will result in spurious results.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES
  • 1
    Perrett DI, Lee KJ, Penton-Voak I et al. Effects of sexual dimorphism on facial attractiveness. Nature 1998; 394: 885 887.
  • 2
    Mikhailova ES, Vladimirova TV, Iznak AF, Tsusulkovskaya EJ, Sushko NV. Abnormal recognition of facial expression of emotions in depressed patients with major depression disorder and schizotypal personality disorder. Biol. Psychiatry 1996; 40: 697 705.
  • 3
    Matsumoto D. Cultural similarities and differences in display rules. Motivation Emotion 1990; 14: 195 214.
  • 4
    Matsumoto D & Ekman P. American-Japanese cultural differences in judgments of facial expressions of emotion. Motivation Emotion 1989; 13: 143 157.
  • 5
    Matsumoto D. American-Japanese cultural differences in the recognition of universal facial expressions. J. Cross-Cultural Psychol. 1992; 23: 72 84.
  • 6
    Shioiri T, Someya T, Helmeste D, Tang SW. Misinterpretation of facial expression: A cross-cultural study. Psychiatry Clin. Neurosci. 1999; 53: 45 50.
  • 7
    Morris JS, Frith CD, Perrett DI et al. A differential neural response in the human amygdala to fearful and happy facial expressions. Nature 1996; 383: 812 815.
  • 8
    Phillips ML, Young AW, Senior C et al. A specific neural substrate for perceiving facial expressions of disgust. Nature 1997; 389: 495 498.
  • 9
    Matsumoto D & Ekman P. Japanese and Caucasian Facial Expressions of Emotion (JACFEE) Slides. Intercultural and Emotion Research Laboratory, Department of Psychology, San Francisco State University, San Francisco, CA, 1988.
  • 10
    Pilowsky I & Katsikitis M. The classification of facial emotions: A computer-based taxonomic approach. J. Affect. Dis. 1994; 30: 61 71.
  • 11
    Russell JA. Facial expressions of emotion: What lies beyond minimal universality? Psychol. Bull. 1995; 118: 379 391.
  • 12
    Russell JA & Bullock M. Fuzzy concepts and the perception of emotion in facial expressions. Social Cognition 1986; 4: 309 341.
  • 13
    Nummenmaa T. The recognition of pure and blended facial expressions of emotion from still photographs. Scand. J.Psychol. 1988; 29: 33 47.