Transcranial direct current stimulation (tDCS) targeting the postcentral gyrus reduces malevolent creative ideation

Abstract Malevolent creativity (MC) is defined as a manifestation in which people propose to materially, mentally or physically harm themselves or others in a novel manner. Malevolent creative ideation can be inhibited by high moral emotions (i.e. sympathy, guilt and shame) and low negative emotions, which promote prosocial behaviors. Given that the right postcentral gyrus (PCG) is involved in generating sympathy and emotional recognition for others and the right middle frontal gyrus (MFG) is involved in emotional regulation, we suggest that the right PCG and right MFG may play important roles in malevolent creative ideation. In Study 1, we recruited 98 healthy and right-handed college participants (80 females, age = 21.11 ± 2.00 years) and examined the role of the right PCG in malevolent creative ideation using transcranial direct current stimulation (tDCS). The results showed that the accuracy of emotional recognition changed when the right PCG received electrical stimulation. Enhancing the activation of the right PCG reduced MC originality and fluency, whereas inhibiting it increased MC originality and fluency. In Study 2, we recruited 91 healthy and right-handed college participants (74 females, age = 21.22 ± 2.28 years) and examined the role of the right MFG in malevolent creative ideation using tDCS. The results showed no significant difference in malevolent creative performance between the pre- and post-test when electrical stimulation was applied over the right MFG. These findings indicate that enhancing the activation of the right PCG, which is closely correlated with emotional recognition, reduces an individual’s malevolent creative ideation.


Introduction
Creativity is considered as the ability to generate novel and useful ideas or products (Runco and Jaeger, 2012). As one of the dark aspects of creativity, malevolent creativity (MC) is defined as a manifestation in which individuals propose to materially, mentally or physically harm themselves or others in a novel manner (Cropley et al., 2008;Cropley, 2010). Based on the definition of MC, it involves not only regular creativity but also a purpose to harm others mentally or physically. This emphasizes that MC is a more complex cognition and intrinsically differs from regular creativity. While MC may be deemed essential for a particular group or individual to fulfill their goals, it has negative consequences for other groups or individuals (Cropley et al., 2008). Examples of MC exist everywhere, such as fraud and creative crimes, and can lead to severe negative consequences. Research focused on the relationship between malevolent creative performance and criminal behavior (e.g. murder, abuse and violence), personality, motivation and moral emotion (Cropley et al., 2008;Harris et al., 2013;Hao et al., 2020;Cheng et al., 2021b). A possible link between MC and criminal behavior can be found when individuals break the law. Cropley et al. (2008) demonstrated that a creative criminal might carry out a crime in an original way. Personality traits such as aggression are correlated with MC. Individuals who show higher physical aggression exhibit higher levels of MC (Lee and Dow, 2011). Hao et al. (2020) investigated the relationship between motivation and MC and found that approach motivation predicted MC, while avoidance motivation was negatively associated with MC.
Moral emotion and MC may be closely linked. Kapoor and Kaufman (2022) proposed the AMORAL model in relation to MC. The AMORAL model consists of the Antecedents, Mechanisms (individual), Operants (environmental), Realization, Aftereffects, and Legacy of dark creativity. Regarding mechanisms (individual), these tend to include certain individual-level features that can inhibit or promote MC. In particular, this individuallevel feature refers to a range of social and emotional abilities, such as identifying and managing emotions (Kapoor and Kaufman, 2022).
On the one hand, identifying emotions is closely associated with malevolent creative performance. Experiencing moral emotions, such as guilt and shame based on identifying emotions, can discourage creative individuals from engaging in more malevolent actions (Haidt, 2003b;Tangney et al., 2007). Guilty conscience, viewed as a superego response to unacceptable impulses (Babcock, 1983), can promote prosocial behaviors (Ferguson and Branscombe, 2010;Rees et al., 2015;Hurst and Sintov, 2022). Shame also leads to higher levels of guilt, which in turn promotes positive behavior (Rees et al., 2015). In addition, sympathy, another moral emotion, and MC may be closely related. Sympathy is defined as an understanding of others' situations and feelings (Eisenberg, 2000;Malti et al., 2013). Sympathy and prosocial/cooperative behavior are generally found to have a positive relationship (Eisenberg and Miller, 1987;Eisenberg et al., 1989;Carlo et al., 2010;Grueneisen and Warneken, 2022). Sympathy can motivate individuals to react sympathetically to others' suffering, and individuals with higher sympathy tend to find novel and prosocial solutions that alleviate suffering and promote well-being (Yang and Yang, 2016), which is the opposite of MC. Unethical behavior is significantly predicted by MC (Gino and Ariely, 2012). When implementing malevolent creative ideas, individuals may not necessarily gain substantial profits, especially when deceiving and retaliating against others. In this case, individuals gain psychological satisfaction by putting others in an unfortunate situation or causing them to feel negative emotions. This suggests that individuals with high levels of MC could predict their actions that will lead to a negative emotional reaction in the recipient. However, they may become less emotionally activated by such actions in order to avoid sympathy, guilt and shame.
On the other hand, managing emotions may also be related to malevolent creative performance. Dark creativity involves the process of malevolent creative ideation (Cropley et al., 2008;Cropley, 2010). Individuals with high MC may have the tendency to be activated with proactive emotions (i.e. anger) in response to other people's doing or in anticipation of a psychological reward from other people's suffering. Anger can explain the unique, non-overlapping variance in the capacity to implement MC (Perchtold-Stefan et al., 2021). Similarly, Cheng et al. (2021a) investigated the effects of anger on MC. The results showed that idea fluency and originality were higher in the group experiencing anger than those in the group experiencing neutral emotion. Cheng et al. (2021b) examined whether emotional regulation could modulate such an effect. They determined that cognitive reappraisal and expression inhibition can lower anger and decrease malevolent creative performance. Accordingly, managing negative emotions may be a useful way of inhibiting MC.
The neural correlates of emotion identification are related to the postcentral gyrus (PCG) (Bufalari et al., 2007;Bertsch et al., 2013;Kropf et al., 2019). The PCG is known to perceive primary sensations and house the secondary somatosensory cortex, which appears to integrate somatosensory stimuli and memory formation (Chen et al., 2008). The PCG forms part of the somatosensory cortex, which is involved in emotional reactions due to empathetic responses (Bufalari et al., 2007). Increased activity in the PCG appears to increase the identification of others' feelings (Bufalari et al., 2007;Kropf et al., 2019). Recognition of others' emotions also requires sensory involvement (Pitkä nen, 2018; Yavuz et al., 2018). In other words, the PCG is involved in the process of recognizing others' emotions by integrating stimuli and sensations as the secondary somatosensory cortex. As mentioned above, recognizing others' emotions is important for MC. In addition, the right hemisphere, especially the PCG, is particularly specialized in processing moral emotions (Borod et al., 1998;Canli, 1999;Yoshino et al., 2017), and these inductions affect malevolent creative performance (Fu and Zhang, 2022;Kapoor and Kaufman, 2022). Previous studies have also suggested that individuals with higher levels of psychopathy have a smaller volume of the PCG (Tiihonen et al., 2008;Martina Ly et al., 2012;Bertsch et al., 2013) and higher aggression (Jonason et al., 2014), which is associated with higher levels of MC (Lee and Dow, 2011). Based on the aforementioned research, it can be inferred that the right PCG is closely related to MC.
Managing emotions is closely correlated with activity in the middle frontal gyrus (MFG) (Ochsner et al., 2004;Ohira et al., 2006;Blair et al., 2007). Previous findings have indicated the importance of the lateral frontal regions in emotional regulation (Ochsner et al., 2004;Blair et al., 2007). Emotional regulation is usually related to the activation of the MFG through suppression (Ohira et al., 2006;Blair et al., 2007). Brain imaging findings have indicated a positive connection between the MFG and the lateral superior pre-frontal gyrus, previously thought to be implicated in emotional management/regulation (Beauregard et al., 2001;Ohira et al., 2006). In particular, the right MFG is involved in downregulation and suppression effects, which are key processes for successful emotional regulation (Grecucci et al., 2012;Engen and Anderson, 2018). As mentioned above, emotional regulation can affect malevolent creative performance. In so doing, the activation of the right MFG may also affect malevolent creative performance.
In the present study, we designed two studies to explore the following questions: does altering the activity of the right PCG and right MFG affect one's emotional recognition and regulation function and does this in turn affect malevolent creative ideation? In Study 1, we assumed that when the activity of the right PCG is inhibited, individuals' malevolent creative performance would be strengthened. If activity in the right PCG is not inhibited, an individual's malevolent creative performance would weaken. In Study 2, we assumed that when the activity of the right MFG is higher, then the individual's malevolent creative performance would be decreased. In contrast, an individual's malevolent creative performance is enhanced when activity in the right MFG is decreased. Here, we altered the activity of the right PCG or right MFG using transcranial direct current stimulation (tDCS). The tDCS technique, a non-invasive brain stimulation method, can modulate cortical excitability. It can deliver a low-intense (0.5-2.0 mA) direct current to specific cortical areas and facilitate (anode) or inhibit (cathode) spontaneous neuronal activity (Nitsche and Paulus, 2000;Brunoni et al., 2012). Research has shown that tDCS can affect creative performance by modulating the excitability of the cortical regions associated with creative thinking (Chi and Snyder, 2012;Mayseless and Shamay-Tsoory, 2015;Weinberger et al., 2017;Hertenstein et al., 2019;Ivancovsky et al., 2019;Kleinmintz et al., 2019). Ivancovsky et al. (2019) investigated the effect of tDCS of the inferior frontal gyrus (IFG) on creative divergent thinking. The results showed that individuals exhibited increased creative performance after receiving cathodal stimulation over the left IFG, whereas individuals exhibited poor creative performance after receiving anodal stimulation over the left IFG. Mayseless and Shamay-Tsoory (2015) found that anodal tDCS over the right IFG, coupled with cathodal tDCS over the left IFG, promoted verbal creative divergent thinking, whereas reverse stimulation did not affect verbal creative divergent thinking. Chi and Snyder (2011) investigated the effect of tDCS on the anterior temporal lobe (ATL) using the matchstick arithmetic task (a creative task) (Öllinger et al., 2008). They found that anodal tDCS over the right ATL, coupled with cathodal tDCS over the left ATL, increased creative performance. In light of this, it was determined that tDCS modulates neuronal activity in brain regions correlated with creativity, which in turn affects creative performance (Weinberger et al., 2017;Chrysikou et al., 2021;Lifshitz-Ben-Basat and Mashal, 2021).

Study 1
We examined whether altering the activity of the right PCG affects emotional recognition and malevolent creative performance using tDCS.

Participants
One hundred and one right-handed college students were randomly assigned to three groups (anodal, cathodal and sham tDCS groups). Participants were recruited using online posters. Data of 98 participants who completed all the tasks (80 females, age = 21.11 ± 2.00 years) were entered into subsequent analysis. The anodal tDCS group (33 participants, 25 females, age = 21.31 ± 2.06 years), cathodal tDCS group (35 participants, 29 females, age = 20.97 ± 1.92 years) and sham tDCS group (30 participants, 26 females, age = 21.03 ± 2.13 years) received positive, negative and sham stimulation over the right PCG, respectively ( Figure 1A). All participants had no history of alcohol or drug abuse, neurological or psychiatric abnormalities, claustrophobia or head injury. All procedures were approved by the University Committee on Human Research Protection (UCHRP) of East China Normal University (ECNU), and all participants provided informed written consent for this study.

Transcranial direct current stimulation
Direct current was delivered through a 1 × 1 tDCS low-intensity stimulator using a pair of saline-soaked electrodes (3 × 3 cm 2 ). In the stimulation groups, participants continually received a constant current of 1.5 mA while solving creative tasks for 20 min. In the sham stimulation group, only a fade-in current (30 s) and a fade-out current (30 s) were successively delivered  to the participants before the task sessions ( Figure 1B). The relevant tDCS electrode was placed over the right PCG, and the non-relevant tDCS electrode was placed over the left mastoid ( Figure 1A).

Experimental protocol
All participants completed all tests within 2 days. The experimental procedure comprised two parts which were, respectively, completed in two consecutive days ( Figure 2A). On Day 1 (Part 1), the participants were asked to complete a pre-test creativity task session, which comprised five 2-min malevolent creative ideation tasks (MCTs) and five 2-min benevolent creative ideation tasks (BCTs) (pre-test of creative tasks). The MCTs and BCTs were adapted from the realistic presented problems task, which asked participants to solve open-ended realistic problems originally (Okuda et al., 1991;Agnoli et al., 2016). For the MCTs, an example is 'your friend, who hates to dress the same clothes as others, found that a classmate has the same coat as his. Please think of a novel manner to destroy his classmate's coat secretly'. For the BCTs, an example is 'your friend is too shy to meet the girl he loves. Please produce a novel manner to help him to get to know the girl naturally.' These five MCTs or BCTs were randomly presented. Participants were required to generate as many novel responses as possible and type them into the computer.
On Day 2 (Part 2), participants completed three emotional self-reports, two emotional recognition tasks and two emotional regulation tasks. Here, the emotional tasks were used to examine whether emotional recognition/regulation was affected by tDCS stimulation. One single trial of the emotional regulation task included a jittered fixation cross (2 s), a negative image (8 s) that could induce negative emotions and a two-item self-report. Participants were required to regulate their negative emotions induced by these images. One of the self-report items was rating negative emotions induced by negative images using a 1-5 scale (1 = not at all; 5 = fully induced). The other was rating the levels of self-regulation using a 1-7 scale (1 = not at all; 7 = fully regulated). The emotional regulation task contained 10 pictures (i.e. 10 trials) from the International Affective Picture System (Bradley and Lang, 2007) ( Figure 2B). The emotional recognition task presented face pictures with different emotions from the Karolinska Directed Emotional Faces (KDEF) to participants, including afraid, angry, disgusted, happy, sad and surprised faces (Goeleven et al., 2008). Four sets of face pictures (i.e. four different faces) were selected. Each set consisted of six pictures referring to the above-mentioned six types of emotions (i.e. a total of 4 × 6 = 24 pictures). The presenting sequence was random among participants. One single trial of the emotional recognition task comprised a jittered fixation cross (1 s), an emotional face picture (8 s) and a choice question. Participants were asked to choose the correct emotion corresponding to the face pictures ( Figure 2C). The respond time and accuracy were recorded for subsequent analysis.
First, participants were asked to complete the first emotional self-report using a 9-point Likert scale (1 = not at all; 9 = fully; ∼1 min). The items included worry, disappointment, anger, anxiety, happiness, satisfaction, stress, discouragement, relaxation, pleasure and activation level. Second, participants were required to complete the two emotional tasks (pre-test of the two emotional tasks), and these tasks were randomly presented to participants (∼8 min). Third, participants completed the second emotional self-report (∼1 min). Then, participants received tDCS and completed the post-test creativity task during electrical stimulation (20 min). Fourth, participants completed the third emotional self-report (∼1 min). Finally, participants completed the two emotional tasks again (post-test; ∼8 min). See details in Figure 2A.

Creative performance scoring
The performance of creativity tasks was assessed using idea fluency, originality and benevolence/malevolence Qiao et al., 2022). Fluency was calculated based on the number of ideas. We assessed the originality, malevolence and benevolence scores by referring to the originality, malevolence and benevolence of each answer to evaluate task performance for each participant. Four trained raters independently assessed the originality, malevolence and benevolence scores for each answer to provide an index of creative, malevolent and benevolent quality using a 1-5 scale (1 = not creative/malevolent/benevolent at all; 5 = very creative/malevolent/benevolent) (Amabile, 1982). The final originality, malevolence and benevolence scores for each participant were obtained by averaging the individual ratings from all items. Individual ratings for each participant from these four raters were averaged into a single originality, malevolence or benevolence score. The Internal Consistency Coefficients (ICCs) of the four trained raters were acceptable (Supplementary Table S1).

Emotional self-report
One-way repeated measures Analysis of Variances (ANOVAs) using TEST (first test vs second test vs third test) as the within-subject factor were performed on worry, disappointment, anger, anxiety, happiness, satisfaction, stress, discouragement, relaxation, pleasure and activation level scores. The details of the results are presented in Table 1.

Emotional task performance
Regarding the emotional recognition tasks, the response time and accuracy were submitted to two-way mix-design ANOVAs with STIMULATION (anodal, cathodal and sham stimulation) as the between-subject factor and TEST (pre-test vs post-test) as the within-subject factor. Results suggested no

Malevolent creative task performance
The fluency, originality and malevolence scores were submitted to two-way mix-design ANOVAs with STIMULATION as the betweensubject factor and TEST as the within-subject factor. Regarding

Benevolent creative task performance
The fluency, originality and benevolence scores were submitted to two-way mix-design ANOVAs with STIMULATION as the betweensubject factor and TEST as the within-subject factor.
We did not find significant main effects on the fluency

Interim discussion
Study 1 aimed to explore whether activity in the right PCG affected the identification of emotions and then affected malevolent creative performance. Results showed significantly lower accuracy in the pre-test than that in the post-test under the sham stimulation condition. The MCT results showed lower fluency scores in the post-test but higher originality scores in the pre-test under the sham stimulation condition. Consistent with previous findings, the results may indicate that individuals may need more time to generate more novel ideas (Wang et al., 2017(Wang et al., , 2019. In MCTs, the originality scores were significantly higher in the pre-test than those in the post-test under the anodal and sham stimulation conditions, whereas lower originality scores were found in the pre-test under the cathodal stimulation condition. Furthermore, MCT results indicated lower fluency scores in the pre-test under the cathodal and sham stimulation conditions, whereas higher fluency scores in the pre-test under the anodal stimulation condition. That is, augmenting the right PCG reduced malevolent creative performance, whereas inhibiting it enhanced individual malevolent creative performance. Meanwhile, electrical stimulation also affected individual emotional recognition ability. In addition, we did not find a significant difference between the pre-and post-test in BCTs. However, another possibility could be that social pressure (feelings of guilt and shame) might activate the PCG during MC, which eventually impaired MC originality. It would be helpful to examine the effect of social pressure on MC performance in further research.

Study 2
In Study 2, we examined whether altering the activity of the right MFG affects emotional regulation and malevolent creative performance using tDCS.

Participants
One hundred right-handed college students were randomly assigned to three groups (anodal, cathodal and sham stimulation groups). Participants were recruited using online posters. Data of 91 participants who completed all the tasks (74 females, age = 21.22 ± 2.28 years) were entered into subsequent analysis. The anodal tDCS group (32 participants, 25 females, age = 21.44 ± 2.14 years), cathodal tDCS group (29 participants, 24 females, age = 21.17 ± 2.74 years) and sham tDCS group (30 participants, 26 females, age = 21.00 ± 2.10 years) received positive, negative and sham stimulation over the right MFG, respectively ( Figure 1C). All participants had no history of alcohol or drug abuse, neurological or psychiatric abnormalities, claustrophobia or head injury. All procedures were approved by the UCHRP of ECNU, and all participants provided informed written consent for this study.

Transcranial direct current stimulation
The tDCS procedure was similar to Study 1. The relevant tDCS electrode was placed over the right MFG, and the non-relevant tDCS electrode was placed over the left mastoid ( Figure 1C).

Experimental protocol
The experimental protocol was similar to Study 1.

Scoring and statistical analysis
Scoring and statistical analysis were similar to Study 1. The ICCs are presented in Supplementary Table S2.

Emotional self-report
One-way repeated measures ANOVAs using TEST (first test vs second test vs third test) as the within-subject factor were performed on worry, disappointment, anger, anxiety, happiness, satisfaction, stress, discouragement, relaxation, pleasure and activation level scores. Results showed no significant main effect of TEST on worry [F (2,176)

Emotional task performance
Regarding emotional recognition tasks, we performed twoway mix-design ANOVAs using STIMULATION (anodal, cathodal and sham stimulations) as the between-subject factor and TEST (pre-test vs post-test) as the within-subject factor on the response time and accuracy. No significant effects were observed ( Figure 4A).
Regarding emotional regulation tasks, the emotional scores and self-regulation scores were submitted to two-way mix-design ANOVAs with STIMULATION as the between-subject factor and TEST as the within-subject factor. Results suggested no significant main effect of TEST [F (1,86)

Malevolent creative task performance
We performed two-way mix-design ANOVAs using STIMULATION as the between-subject factor and TEST as the within-subject factor on the fluency, originality and malevolence scores. Results showed no significant main effects of STIMULATION [F (2,86)

Benevolent creative task performance
The fluency, originality and benevolence scores were submitted to two-way mix-design ANOVAs with STIMULATION as the between-subject factor and TEST as the within-subject factor. Results showed no significant main effects of SIMULATION on the fluency [F (2,86) = 0.08, P = 0.92], originality [F (2,86) = 2.57, P = 0.08] Fig. 4. A significant difference between the pre-and post-test for emotional recognition, emotional regulation, malevolent creative and benevolent creative tasks in Study 2. tDCS was applied over their right MFG and left mastoid. (A) A significant difference between the pre-and post-test for emotional recognition tasks. (B) A significant difference between the pre-and post-test for emotional regulation tasks. (C) A significant difference between the pre-and post-test for malevolent creative tasks. (D) A significant difference between the pre-and post-test for benevolent creative tasks. *P < 0.05; *** P < 0.0001.

Interim discussion
In Study 2, we aimed to explore whether inhibiting activity in the right MFG may decline emotional regulation but enhanced individuals' malevolent creative performance, whereas enhancing it may lead to a decline in malevolent creative performance. We found a significant difference between the pre-and post-test on emotional regulation. However, in the creative tasks, results showed no significant interaction effect of TEST × STIMULATION on the fluency, originality and malevolence/benevolence scores. Our results were inconsistent with our hypothesis. Although inhibiting activity in the right MFG may decline emotional regulation, individuals' creative performance was not affected. We assumed that the right MFG was involved in the process of emotional regulation. Nevertheless, the right MFG is involved in many other high-level cognitive functions such as working memory (Mottaghy et al., 2003;Jolles et al., 2013;Subramaniam et al., 2014) and attention (Vossel et al., 2014;Agmon et al., 2021), which are closely associated with creativity (Vartanian, 2009;Benedek et al., 2014;Zabelina, 2018). From this point of view, it seemed that if activity in the MFG was enhanced, creative performance should also be enhanced, as this result was contrary to our hypothesis. We speculated that augmenting activity in the MFG enhanced individuals' emotional regulation but reduced malevolent creative performance. Under the circumstances, individuals' malevolent creative performance was not changed because the right MFG was involved in both emotional regulation and creative cognition processing.

Discussion
This study explored the relationship between malevolent creative performance and the activity of the right PCG or right MFG using the tDCS technique. Study 1 showed that augmenting the right PCG using tDCS decreased the performance of MC, whereas inhibiting it enhanced the performance of MC. Study 2 showed that stimulating the right MFG with tDCS did not affect an individual's performance of MC.
Study 1 showed that augmenting the activity of the right PCG impaired individual malevolent creative performance, whereas inhibiting this brain region enhanced malevolent creative performance. Previous research has indicated that the right PCG is involved in the identification of others' emotional expressions (Bufalari et al., 2007;Morawetz et al., 2020). The primary and secondary somatosensory cortices have a closed relation in facial emotional recognition, helping individuals construct somatosensory images related to emotions (Adolphs et al., 2000;Adolphs, 2002). As one of the primary and secondary somatosensory cortices, the PCG is also involved in emotional recognition (Satoh et al., 2002;Bufalari et al., 2007;Chen et al., 2008). Study 1 showed significantly lower accuracy in the emotional recognition task in the pre-test than in the post-test under the sham stimulation condition, possibly owing to the practice effect. However, intriguingly, stimulating the right PCG seemed to counteract the supposed practice effect in the other two conditions. In other words, the practice effect vanished under both the anodal and cathodal stimulation conditions, and even lower accuracy was observed in the post-test than in the pre-test under the cathodal stimulation condition. We did not find a significant difference between the preand post-test in emotional recognition tasks. One possible reason may be due to the face pictures that we chose for the tasks. The face pictures that we chose may need to be more suitable for the tasks because the face pictures were randomly selected from the KDEF, and we did not check the accuracy of emotional recognition in Chinese. Choosing more suitable stimulus materials is a possible solution for future exploration.
The AMORAL model proposes that socio-emotional skills, such as emotional recognition, are important contributing factors to MC (Kapoor and Kaufman, 2022). Moral emotions (i.e. sympathy and guilt) can promote prosocial behaviors (Ferguson and Branscombe, 2010;Rees et al., 2015;Hurst and Sintov, 2022). Sympathy, defined as understanding and sharing of another individual's feelings, is associated with emotional recognition. Sympathy leads individuals to react sympathetically to others' suffering, and individuals with higher sympathy tend to find novel and positive solutions to promote prosocial behavior (Rees et al., 2015;Yang and Yang, 2016). Sympathy may interfere with malevolent creative performance (Eisenberg, 2000;Malti et al., 2013). Similarly, recognizing others' emotional expressions or sensations, which is linked to the activity of the right PCG (Adolphs et al., 2000;Bufalari et al., 2007), may cause specific negative moral emotions (i.e. guilt and shame) and thus prevent malevolent actions. Individuals are less likely to create malevolent consequences when experiencing guilt and shame (Rees et al., 2015;Kapoor and Kaufman, 2022). Consistent with these findings, Perchtold-Stefan et al. (2022) found that higher levels of MC corresponded with relative increases in electroencephalogram coherence during others' expressions of anger. Individuals recognized others' anger with greater emotional detachment, possibly indicating that they were unperturbed by the potential consequences of their malevolent actions (Perchtold-Stefan et al., 2022). This combination suggests that malevolent creative performance may be affected by the identification of others' emotions. Augmenting the right PCG weakens malevolent creative performance, most likely through enhancing the identification of others' emotions. Although no significant difference was observed in emotional recognition tasks under electrical stimulation conditions, it was possible that MC was still affected due to negative social emotions, which referred to provide immediate and salient feedback on our social and moral acceptability, such as guilt, shame and sympathy (Haidt, 2003a;Tangney et al., 2007). Meanwhile, such emotions also provide the motivational force to do good and avoid doing bad (Kroll and Egan, 2004). However, this study did not measure these social emotions. We should examine emotional recognition ability directly and corresponding changes of social emotions in subsequent research.
Although the results showed that malevolent creative performance decreased significantly when the activity of the right PCG was enhanced, there was no significant difference in the pre-test and post-test of benevolent creative performance. This may indicate that the activity of the right PCG, which is closely linked to emotional recognition (Bufalari et al., 2007), had a greater influence on MC than on other forms of creativity. However, this influence needs to be further examined in future studies.
The AMORAL model also proposed that emotional regulation is closely interlinked with MC (Kapoor and Kaufman, 2022). Study 2 explored whether inhibiting activity in the right MFG reduced emotional regulation and, as a consequence, increased malevolent creative performance. Although we found a significant difference between the pre-and post-test in the emotional regulation tasks, results of the creative tasks showed no significant difference between the pre-and post-test for fluency, originality and malevolence/benevolence. These results are inconsistent with our hypotheses. That is, although inhibiting the activity of the right MFG may reduce emotional regulation, individuals' creative performance is not affected.
The right MFG is involved in emotional regulation. Augmenting the activity of the MFG enhances individuals' emotion regulation ability, and individuals may experience more negative emotions in the process of malevolent creative ideation. Previous findings indicate that negative feelings affect MC (Cheng et al., 2021a,b). For example, Cheng et al. (2021a,b) found that fluency and originality were higher in the anger group than those in the neutral group and that cognitive reappraisal and expression inhibition can decrease malevolent creative performance. Negative emotions also lead to the experience of guilt, thus promoting positive behaviors (Rees et al., 2015), which are not conducive to malevolent creative ideation. Given this, managing negative emotions may reduce malevolent creative performance. Nevertheless, the right MFG is involved in many other high-level cognitive functions such as working memory (Jolles et al., 2013;Subramaniam et al., 2014) and attention (Vossel et al., 2014;Agmon et al., 2021). Such cognitive functions are closely related to creativity (Benedek et al., 2014;Zabelina, 2018). Accordingly, it seems that if MFG activity was enhanced, creative performance should also be enhanced, which is contrary to the influence of emotional regulation on malevolent creative performance. Under these circumstances, an individual's malevolent creative performance does not change because the right MFG is involved in both emotional regulation and creative cognitive processing. In addition, the right MFG is less active under certain emotional states. It is possible that the emotional tasks preceding the electrical stimulation and creativity task created a specific state-dependent brain response. In this case, the activation or deactivation of the right MFG did not affect MC performance (Bogdanov and Schwabe, 2016).
Some limitations of this study warrant further discussion. First, only college students were recruited as participants; as such, it is necessary to further examine whether these findings can be generalized to other groups. Second, given that the right MFG is also involved in a variety of high-level cognitive functions, whether stimulating the right MFG affects other cognitive functions should be further examined. Third, although we stimulated the target brain regions using tDCS, it was unclear whether the activity of the target brain regions was actually enhanced or inhibited after receiving electrical stimulation. Therefore, it would be helpful to utilize several neuroimaging techniques (e.g. functional magnetic resonance imaging and functional near-infrared spectroscopy) to assess the activity of the target brain regions after stimulation in future research.

Conclusion
Using the tDCS technique, we found when enhancing the activity of the right PCG reduced malevolent creative performance, it was possible due to individuals' emotional recognition improved. That is, individuals were affected by sympathy, guilt and shame; thus, their malevolent creative performance would be decreased. In a word, malevolent creative performance may be affected by the identification of others' emotions, which was relatively associated with the activity of the right PCG.

Supplementary data
Supplementary data are available at SCAN online.

Data availability
The data and code used to support the findings of this study are available from the corresponding author upon request. The data can only be used for research. If the associated research is to be published, the statement 'The data and code were acquired from the Shanghai Key Laboratory of Mental Health and Psychological Crisis Intervention, School of Psychology and Cognitive Science, East China Normal University' is required in the manuscript'.

Author contributions
Z.G. and N.H. conceived the experiment. Z.G. performed the research. Z.G. analyzed the data. Z.G. and K.L. visualized the results. Z.G., K.L. and N.H. wrote the paper.

Funding
This work was sponsored by the National Natural Science Foundation of China (31971002) to N.H.