Dimensional Classification Of Facial Expressions Of Emotion On Intensity Ratings

[Background] Matsumoto and Ekman's Japanese and Caucasian facial expressions of emotion (JACFEE) have been tried in many countries including China. Consistent results from these nominal (or categorical) studies have been shown that the perception accuracies of Anger, Disgust, Happiness and Sadness are high, and that of Contempt is low, and those of Fear and Surprise are mixed. We have hypothesized that a principal component analysis (PCA) on the perceived intensity alone might offer a scalar (dimensional) classification of emotion.[Methods] We invited 200 Chinese university students who were blind to the JACFEE nominal classification, to rate the emotion intensities displayed by each poser. To combine intensity ratings, mean ratings were calculated across the expressions labeled as anger, contempt, disgust, fear, happiness, sadness and surprise, respectively. Gender differences of the rating scores were evaluated by one-way ANOVA. The intensity rating scores of the 56 photos by individual judges were analyzed by a PCA.[Results] No significant between-gender difference was found on intensity rating to each facial expression of emotion. Happiness received the highest intensity score, and Contempt the lowest. PCA results produced seven scalar factors of emotion which corresponded well to the nominal ones. Picture loadings on the target emotions demonstrated that Contempt emerged as an independent factor, similarly to Anger, Disgust, Happiness and Sadness. However, picture loadings in regard to Fear and Surprise were mixed up, which still leaves the question open how to separate the two emotions efficiently.[Conclusion] Our study has been demonstrated the validity of JACFEE in Chinese culture, and suggests a modification of the picture clusters, especially those of Fear and Surprise.