**2.2 Method**

#### **2.2.1 Participants**

Thirty-seven healthy volunteers (18 females; mean age ± SD: 19.5 ± 3.4) were divided into two groups according to the perspective taking scale of a Japanese version of the IRI (Davis, 1983): 19 participants with a high perspective-taking ability (mean score 24.79 ± 1.76) and 18 participants with a low perspective-taking ability (mean score 18.83 ± 1.89).

#### **2.2.2 Stimuli**

Pictures of happy and anger facial expressions were made by a morphing technique, based on digitized grayscale images of 6 Japanese faces (3 men and 3 women) showing a neutral

Using NIRS to Investigate Social Relationship in Empathic Process 73

First, the orienting task confirmed that the operation of the social relationship between the participants and the opponents succeeded. The rating scores of "cooperativeness" showed that cooperative others were rated as more cooperative than uncooperative others and neutral others, while uncooperative others were rated as more uncooperative than cooperative others and neutral others. In the rating scores of "likeness," it was confirmed that the cooperative others were liked, whereas the uncooperative others were disliked.

Figure 1 shows behavioral data of the pleasantness points in the perspective taking task. For the pleasantness points, a 2 × 3 × 3 (perspective taking, social relationship, and facial expression) analysis of variance (ANOVA) showed main effects of the three factors. The analysis also showed a significant 3-way interaction. The results showed that taking the other's perspective yielded no differences between high and low abilities of perspective taking. This indicated that the ability of perspective taking was not reflected in the

Further analyses of the interaction indicated that for the cooperative others in both the selfand other-perspective taking, the mean score in the happy condition was higher than that in the neutral and anger conditions, and the mean score in the neutral condition was higher than that in the anger condition (all *p <* .05). However, for the uncooperative others, these differences between facial expressions were only significant in the other-perspective taking. Note that in the self-perspective taking condition to uncooperative others, there was no

Fig. 1. Behavioral evaluation of the validity of the mean value of each condition on a 7-point scale ranging from "pleasantness" to "unpleasantness." Adapted from Nomura et al. (2010).

difference among the rating scores of the three facial expressions.

**2.3 Results and discussion** 

**2.3.1 Behavioral data** 

behavioral data.

facial expression. Participants in a pilot study rated each facial expression on a 5-level scale (0 = not at all to 4 = very intense) in terms of happiness and anger. The mean rating for the valence of the selected stimuli was 2.32 ± 0.73 for anger and 2.57 ± 0.89 for happiness. Furthermore, the intensity of each facial expression was rated and the rating scores for the intensity of all selected facial stimuli showing anger and happiness were 3.40 ± 0.92 and 3.15 ± 0.86, respectively.
