The interaction between pain and attractiveness perception in others

When considering the “beauty-is-good” stereotype, facial attractiveness should facilitate empathy for pain. On the other hand, having in mind the “threat value of pain” hypothesis, pain cues would be more salient, and thus, its processing would not suffer influence by facial attractiveness. The event-related potential (ERP) allows investigating if one of these theories could predict individuals’ responses regarding the perception of pain or attractiveness in others’ faces. We tracked 35 participants’ reactions to pictures depicting more and less attractive faces displayed in a painful and non-painful condition. Each participant completed the following two tasks when presented the images of faces: (1) the Pain Judgment Task, in which participants should rate the pain levels, and (2) the Attractiveness Judgment Task, in which participants should rate the attractiveness. Results showed that participants exhibited differences rating more and less attractive faces in the non-painful pictures, but not in the painful pictures. These results were observed in P3 and LPC amplitudes in the Pain Judgment Task, as well as in N170 and P2 amplitudes in the Attractive Judgment Task. Our results suggested that both explicit and implicit empathic pain processing inhibited the processing of attractiveness perception. These findings supported the “threat value of pain” hypothesis. Besides, in the Attractive Judgment Task, the N170 and P2 amplitudes for more attractive painful pictures were larger than those for more attractive non-painful pictures. In contrast, no significant difference was found between the amplitudes for painful and non-painful, less attractive pictures. Our findings suggest that explicit facial attractiveness processing for more attractive face images potentiates the implicit empathy for pain processing, therefore partly supporting the “beautiful-is-good” stereotype.

emotions in the faces, such that negative emotions suppressed attractiveness assessments as compared to positive and neutral emotional faces 19,20 . According to the "threat value of pain" hypothesis, others' pain, serving as warning signals to avoid or escape [5][6][7][8] , would elicit negative emotional reactions [21][22][23][24] , thus suppressing the perception of physical attractiveness, as evidenced by decreased attractiveness ratings and neural responses.
Both others' pain (signaling threat and danger) and attractiveness (signaling fertility and health) could be processed prioritized by the human attention system 25,26 . The present study aimed to explore the interplay between others' pain and physical attractiveness as perceived by participants. More specifically, we examined (1) whether empathy for pain can be influenced by others' attractiveness and (2) whether perceptions of others' attractiveness can be influenced by others' pain. Therefore, four categories of facial stimuli were adopted, i.e., non-painful and painful facial stimuli with either low or high physical attractiveness. Each participant completed two tasks. In the first, we applied the Pain Judgment Task, in which participants were asked to judge whether facial stimulation was painful or non-painful. In this way, others' pain would be processed explicitly, while attractiveness would be processed implicitly. The second task consisted of the Attractiveness Judgment Task, in which participants were asked to assess whether the presented face was attractive or non-attractive. In this task, attractiveness processing would be explicit processed and pain processing would be implicit.
As predicted by the "beauty-is-good" stereotype, more attractive faces are perceived as being morally good and as having better personalities 13 , thus capturing more attention resources. Therefore, we hypothesized that both of the explicit and implicit processing of others' pain would be potentiated by physical attractiveness processing, as reflected by heightened empathic neural responses to faces with high physical attractiveness. As predicted by the "threat value of pain" hypothesis, the perception of others' pain and injure is a potential threat to the self, apparently provoking observers' threat-detection system and possibly activating a general aversive response 7 . Therefore, we hypothesized that both of the explicit and implicit processing of others' attractiveness would be suppressed by processing of others' pain, as reflected by the decreased processing of facial attractiveness caused by pain cues present in the faces.

Pain judgment task.
As shown in the top panel of Fig. 1, we did not observe a significant main effect or interaction effect in RTs and ACCs (p > 0.05 for all comparisons). Grand average ERP waveforms to painful and non-painful pictures with high and low attractiveness in the Pain Judgment Task, as well as scalp topographies of dominant waves, are shown in Fig. 2. Regardless of whether painful or attractive, others' faces elicited N1 and N2 waves over frontal-central electrodes (e.g., FCz and Fz), N170 waves over occipito-temporal electrodes (e.g., PO7 and PO8) and P2, P3 and LPC waves at central-parietal electrodes (e.g., CPz and Pz). Amplitudes of dominant waves in different conditions were compared using a repeated measures two-way ANOVA with factors of "attractiveness" (more vs. less) and "pain" (painful vs. non-painful), and relevant results have been summarized in Table 1.

Discussion
This study investigated interactions between the perception of pain and attractiveness in facial images. Our results showed that neither the explicit or the implicit empathic pain processing suffered influence by attractiveness, thus supporting the "threat value of pain" hypothesis 7 . Also, the explicit facial attractiveness processing for more attractive faces enhanced the implicit empathy for pain processing, therefore, partly supporting the "beautiful-is-good" stereotype 13 .
Pain judgment task. In line with previous studies showing that painful stimuli elicited larger amplitudes relative to non-painful stimuli in the N2 time window 9,27 , the present study showed enlarged N2 amplitudes to painful pictures relative to non-painful pictures. The N2 amplitudes to others' painful faces were suggested to be related to affective components of empathy for pain 28 . Thus, our findings suggest that emotional resonance was elicited to others' facial pain in the present study.
We observed significant P3 amplitude differences between painful and non-painful pictures of less attractive models, which is consistent with findings of previous ERP studies on empathy for pain. Those studies have found larger amplitudes in the processing of painful images compared to non-painful, using images of hands and feet (see review 1 ), and faces 29 . Interestingly, this effect was not observed when assessing more attractive faces. One possible explanation for those findings would be that the implicit attractiveness processing of more attractive faces could attenuate the empathic responses to others' pain. Therefore, when individuals pay attention to the pain cues, empathy for pain could be reduced towards people with more attractive faces. Previous studies had found that empathic responses to others' facial pain might be sensitive to some physical features, for instance, when participants assessed images of faces ethnically different than themselves, their empathy for pain was reduced 29,30 . The same happened regarding tasks involving trustworthy and untrustworthy faces 31 . It might be possible that own-race images or trustworthy faces are perceived as closer to oneself. Previous work has demonstrated that the perceived closeness to another individual enhances empathic responses 32 . Somehow, the more attractive models may be perceived as non-similar to the participants and the facial features may be more or less prioritized based on the implicit-explicit processing line of reasoning 33 . Thus, participants may exhibit inhibited empathic responses to more attractive models' pain when they instructed to pay attention to the pain cues in the stimuli. However, we acknowledge that alternative explanations for these findings may be possible. As more attractive faces elicited larger P3 amplitudes than less attractive faces in later components 34 . Thus, enlarged P3 amplitudes for more attractive non-painful faces decreased the difference between painful and non-painful more attractive pictures. Above all, the results of the Pain Judgment Task suggest that explicit empathic processing to others' facial pain inhibits the implicit attractiveness processing, which supports the "threat value of pain" hypothesis.
Most importantly, in the Pain Judgment Task, P3 and LPC amplitudes were significantly modulated by the interaction between "attractiveness" and "pain". Differences between more and less attractive faces were only found in the non-painful pictures. Both behavioral results (ACCs, RTs, and pain intensity ratings) and ERP amplitudes to painful pictures were not influenced by the implicit attractiveness processing. Hence, these results  www.nature.com/scientificreports www.nature.com/scientificreports/ exhibited decreased accuracy in judging the more attractive painful pictures, suggesting that pain cues decrease the perception of attractiveness, especially to more attractive faces. In addition, attractiveness ratings to the more attractive faces showed a decrease in painful situations than did the less attractive models. These results indicated that pain decreased perceptions of attractiveness on a behavioral level.
In line with previous ERP studies 34,35 , N170, P2, and LPC amplitudes were modulated by "attractiveness", with more attractive faces eliciting a negative deflection in N170 and P2 components and a positive deflection in LPC components relative to less attractive faces. Similarly to previous studies 9,27,29 , we also found a main effect of "pain" in N2 and P3 amplitude in the Attractive Judgment task, with painful pictures eliciting more negative N2 and more positive P3 waves than non-painful pictures. These results may suggest that N2 and P3 are sensitive to others' pain cues, independently of task demands. As frontal N2 to others' painful faces were suggested in relation to affective components of empathy for pain 28 , and as P3 over the posterior parietal cortical area have been thought to link to a cognitive evaluation component of empathy 29 , it appears that processing resources of evaluation of others' pain were recruited automatically in these time windows even though the pain cues in the Attractive Judgment Task were unrelated to task goals.
More importantly, N170 and P2 components were modulated by the interaction between "pain" and "attractiveness"; that is, for the non-painful pictures, N170 and P2 waves to the more attractive faces were more negative than to the less attractive faces, whereas no difference was found between the more and less attractive faces for the painful pictures. Previous ERP findings 27,31 suggested that the N170 and P2 components were involved in the automatic coding of both facial identity and emotional state, such as pain. Thus, our results suggested that implicit pain processing decreased the distinguishing automatic coding process of others' attractiveness in these time windows, supporting the "threat value of pain" hypothesis 7 .
Interestingly, another aspect of the comparison between "pain" and "attractiveness" revealed that for more attractive faces, N170 and P2 amplitudes of painful pictures were larger than non-painful pictures, whereas no significant difference was observed between the painful and non-painful pictures for less attractive faces. These results suggest that explicit facial attractiveness processing for more attractive faces enhanced the implicit empathy for pain. Based on the "beauty-is-good" stereotype 13 , more attractive faces are perceived as being morally good and as having better personalities, thus, this would increase the observers' empathy for pain. Under this perspective, our findings partly supported the "beauty-is-good" stereotype 13 .
Interplay between processing others' pain and attractiveness. The findings of both tasks in the present study indicate that empathy for pain and facial attractiveness may not be processed independently of each other. The perception of facial attractiveness could be suppressed by empathy for pain, regardless of the task, which strongly supports the "threat value of pain" hypothesis 7 . That is, both attractive faces 34,35 and pain cues 36 automatically capture our attention faster and hold it longer. When offered simultaneously, however, individuals would automatically allot more attention to pain cues. On the other hand, the explicit facial attractiveness processing for more attractive faces enhanced the implicit empathy for pain processing. These results partially support the "beauty-is-good" stereotype 13 .
Despite possible implications, several limitations of the present study should be addressed. Firstly, all face images were gray scale: whether these responses related to daily life requires further investigation. Secondly, both female and male prototypes were involved in the study: it may be that the effect is different for male and female faces. Future research should include gender in the experimental design.

Conclusions
To investigate the association between empathy for pain and the attractiveness, this study employed the Pain Judgment Task and the Attractiveness Judgment Task use ERPs. The results suggested that both explicit and implicit empathic pain processing inhibited the processing of others' attractiveness, thus supporting the "threat value of pain" hypothesis 7 . Furthermore, explicit facial attractiveness processing for more attractive faces enhanced the implicit empathy for pain processing, which partly supported the "beautiful-is-good" stereotype 13 .

Materials and methods
Participants. Thirty-five adults (18 females) from the Chongqing Normal University, Chongqing, China, participated in this study as paid volunteers. None of the participants had been previously diagnosed with a psychiatric, medical, or neurological disorder. All participants were right-handed, aged 18-24 years (Mean = 20.7 years, SD = 2.5 years), and in possession of normal or corrected-to-normal vision. All participants signed informed consent after receiving a complete description of the study. All participants gave their free and informed consent to the study before the experiment in accordance with the Declaration of Helsinki, and all procedures were approved by the Chongqing Normal University research ethics committee. The procedures were performed in accordance with ethical guidelines and regulations.
Stimuli. The stimuli were 120 digital pictures of faces depicting painful and non-painful conditions, revised from a picture database that had been previously validated and used in published studies, in which the face images were morphed pictures 37,38 . The database comprised photos of 30 more attractive faces (15 female faces and 15 male faces) and 30 less attractive faces (15 female faces and 15 male faces). Each face was transformed to depict pain by penetrating the model's cheek with a syringe needle, and a non-painful version was created by touching each model's face with a soft object (Q-tip), using the software "Adobe Photoshop CS6" (Fig. 4). Luminance, contrast, and colour were matched between both groups of pictures.
Experimental procedure. The participants were seated in a quiet room with an ambient temperature of about 20 °C. They were instructed to participate in two experimental tasks: (1) the Pain Judgment Task and (2) the Attractiveness Judgment Task. The order of these two tasks was counterbalanced between participants. For both tasks, the order of stimulus presentation was randomized, using the E-Prime (3.0) program. Electroencephalography (EEG) data were recorded during these two tasks.
In the Pain Judgment Task (shown in the left column in Fig. 5), participants were instructed to determine whether the model was experiencing pain. At the start of a Pain Judgment Task trial, a 500 ms fixation cross was presented on a black screen, followed 800-1,500 ms later by a picture, and the participants were instructed to respond as accurately and as quickly as possible by pressing a key (either "1" or "2") to judge whether the picture depicted pain. The picture disappeared from the screen as soon as the participant responded. Key-pressing was counterbalanced across participants to control for order effects. The Pain Judgment Task comprised four blocks with 150 trials per block and an inter-trial interval of 1,000 ms. Each picture was presented five times during this task.
In the Attractiveness Judgment Task (shown in the right column in Fig. 5), participants were instructed to press a key ('1' or '2'), as accurately and as quickly as possible, to select whether the face was more attractive or less attractive. Except for different experimental instructions, procedures and stimuli in this task were identical to those in the Pain Judgment Task.
Measurement of subjective reports. After the EEG recording session (Pain Judgment Task and Attractiveness Judgment Task), participants were instructed to rate the picture attributes of pain intensity (1 = no  Statistical analysis. Subjective ratings of pictures, including ratings of pain intensity, attractiveness and subjective emotional reaction, were compared using repeated measures analysis of variance (ANOVA) with factors of "attractiveness" (more attractive vs. less attractive) and "pain" (painful vs. non-painful). If any main effect or interaction effect was observed, post hoc comparisons were performed.
Behavioral data, including reaction times (RTs) and accuracies (ACCs), and electrophysiological data (peak latencies and amplitudes of dominant ERP components) were compared separately for the Pain Judgment task and for the Attractiveness Judgment task. Subjective ratings to stimuli, behavioral data and electrophysiological data were compared via a repeated measures ANOVA using the within-participants factors, "attractiveness" (more attractive vs. less attractive) and "pain" (painful vs. non-painful). If any main effect or interaction effect was observed, post hoc comparisons were performed.
In addition, a separate analysis with a factor for "order" (Pain Judgment Task First vs. Pain Judgment Task Second) was taken into account the possible order effect with which the tasks were performed. However, none significant comparisons were found in all dependent variables (all p > 0.05). Relevant results have been summarized in Supplementary materials, S. Table 2.
Ethics approval and consent to participate. This research was approved by the Chongqing Normal University research ethics committee. All participants had signed informed consent after being given a complete description of the study. The ethics committee approved this consent procedure.

Data availability
Supplementary data associated with this article can be found in the online version at https://pan.baidu.com/ s/1ndSIVBg-gp_MiyCiqCmEfg.