Medial Prefrontal Cortex Activation Is Commonly Invoked by Reputation of Self and Romantic Partners

The reputation of others influences partner selection in human cooperative behaviors through verbal reputation representation. Although the way in which humans represent the verbal reputations of others is a pivotal issue for social neuroscience, the neural correlates underlying the representation of verbal reputations of others are unclear. Humans primarily depend on self-evaluation when assessing reputation of self. Likewise, humans might primarily depend on self-evaluation of others when representing their reputation. As interaction promotes the formation of more nuanced, individualized impressions of an interaction partner, humans tend to form self-evaluations of persons with whom they are intimate in their daily life. Thus, we hypothesized that the representation of reputation of others is modulated by intimacy due to one’s own evaluation formation of that person. To test this hypothesis, we conducted a functional magnetic resonance imaging experiment with 11 pairs of romantic partners while they viewed an evaluation of a target person (self, partner [intimate other], or stranger [non-intimate other]), made by other evaluators. When compared with strangers, viewing evaluations of self and partner activated overlapping regions in the medial prefrontal cortex. Verbal reputation of self-specific activation was found in the precuneus, which represents self-related processing. The data suggest that midline structures represent reputation of self. In addition, intimacy-modulated activation in the medial prefrontal cortex suggests that the verbal reputation of intimate others is represented similarly to reputation of self. These results suggest that the reputation representation in the medial prefrontal cortex is engaged by verbal reputation of self and intimate others stemming from both own and other evaluators’ judgments.


Introduction
Reputation plays a key role in human society. Reputation can emerge when information about an actor's behavior in one of his or her relationships spreads to other partners via an information network [1]. Regarding cooperation behaviors, humans pay attention to others' reputation for selecting cooperation partners [2,3]. Humans often rely on the verbal reputation of others for social behaviors related to the evaluation of products [4] or people [5]. As a verbal reputation mechanism is unique to humans, processing of the verbal reputation of others is a pivotal issue in social neuroscience.
Several neuroimaging studies have investigated the neural mechanisms underlying reputation processing. Having a positive reputation, which covaries with high social desirable traits, activates the medial prefrontal cortex (mPFC) and striatum [6]. By contrast, a negative reputation activates the mPFC, anterior insula, and anterior cingulate cortex (ACC) [7]. Given these findings, reputation is commonly represented by mPFC activation [8]. Izuma et al. reported that evaluation of a stranger's reputation did not activate the mPFC [6]. However, as we select cooperation partners based on their reputation [2,3], we are able to imagineand form a representation of -the reputation of others. Thus, it is still unclear how reputation of others invokes reputation representation in the mPFC.
In everyday life, an individual can assess reputation of self by using the responses of others to infer how they think about the individual [9,10]. This kind of inference is called reflected appraisal [11,12]. In this sense, reflected appraisal has similar characteristics to reputation representation. Reflected appraisal is constructed using meta-representations of others' evaluations of a target person [13], and reflected appraisal is primarily dependent on one's own evaluation [11]. To represent a reputation, humans might use information from both their own evaluations and others' evaluations, including verbal reputation evaluations made by others. In this type of cognition, one might take an anchoring and adjustment approach, which is defined as adjusting beliefs based on others' viewpoints from an anchor point based on one's own beliefs [14]. In this sense, one's own evaluation, together with others' evaluations, plays a key role in processing reputation. Previous results showing inactivation of the mPFC when processing the reputation of others vs. processing reputation of self [6] could be due to a lack of one's own evaluation. Therefore, we expected that one's own evaluation of the reputation target is utilized to represent the reputation of others. As the motivation for interaction with others promotes the formation of more nuanced, individualized impressions of the interaction target [15], humans tend to form self-evaluations of intimate persons in their daily life. Thus, we hypothesized that activity in the mPFC should increase when reputation information about oneself or an intimate other is being processed, but not when the same information about a stranger is being processed.
To test this hypothesis, we conducted a functional magnetic resonance imaging (fMRI) study with 11 pairs of romantic partners during an adjective judgment task using two 3T MR scanners. Romantic partners were used to gauge the intimacy effect, because romantic love is accompanied by the desire to share emotions and experiences [16], which leads to intimate relationships. Therefore, humans have formed their own detailed evaluation of their romantic partners. In the experiment, each subject viewed photos of three types of evaluation target (self, partner, or stranger), together with an evaluation word ostensibly reflecting the rating of the target. At that time, they were required to judge the social desirability of the rating word irrespective of the target photo.

Subjects
A total of 11 romantic couples took part in the experiment. The mean 6 standard error of the mean (SEM) age was 21.360.4 years (22.360.7 for males and 20.660.5 for females). All subjects had normal or corrected-to-normal visual acuity and were righthanded according to the Edinburgh Handedness Inventory [17] except two female subjects. Subjects were provided with monetary compensation. The study protocol was approved by the ethical committee of the National Institute for Physiological Sciences (Okazaki, Japan). The experiments were undertaken in compliance with national legislation and the Code of Ethical Principles for Medical Research Involving Human Subjects of the World Medical Association (Declaration of Helsinki). All the subjects provided written informed consent. Six subjects were excluded from the reported results (data from 16 subjects are presented here), because one subject reported that they did not believe that the impressions were evaluated by others after the fMRI experiment, four subjects had too much movement (.2 mm) during a run, and the sixth subject had poor task performance (,95% button-press ratio).

Apparatus for Visual Stimulus Presentation
Visual stimuli were presented using Presentation software 14.4 (Neurobehavioral Systems, Inc., California) implemented on a personal computer (dc7900; Hewlett-Packard Japan, Ltd., Tokyo). A liquid crystal display (LCD) projector (CP-SX12000; Hitachi, Ltd., Tokyo) located outside and behind the scanner projected the stimuli through a waveguide to a translucent screen, which the subjects viewed via a mirror attached to the bed of the MRI scanner. The spatial resolution of the projector was 1,0246768 pixels, with a 60-Hz refresh rate. The distance between the screen and each subject's eyes was approximately 175 cm, and the visual angle was 13.8u (horizontal)610.4u (vertical). Responses were collected via an optical button box (HHSC-262; Current Designs, Inc., Philadelphia).

Task Design
The couples participated in two-day sessions, similar to a previous study [6]. The average (6SEM) time between the first and the second experimental days was 10.2762.22 days. On the first day, the couples took part separately in a self-introduction session, during which they initially completed a self-introduction sheet comprising open-ended questions such as ''What do you do in your free time?'', ''What is your personality like?'', ''What are your goals for the future?'', and ''Please pick one problem that modern Japanese society faces and briefly state your opinion for tackling the issue''. After completing the self-introduction sheet, they were required to tell a self-introduction story, which was video recorded. At the beginning of the session, they were told that the information provided would be used by four couples (two groups of four evaluators) to form an impression of the participant. At the first session, we confirmed that the evaluators were strangers to the participants.
On the second day, the subjects took part in the fMRI experiment during which they were presented with the results of the impression evaluations (adjectives) with a photo of the participant (self), his/her partner, or a stranger. The adjectives were presented in a predetermined order. We selected 68 adjectives from 84 items used in a previous study [6]. The adjectives were selected based on the rating results of nine independent evaluators (five males) using a seven-point Likert scale (ranging from 7, highly understandable to 1, not at all understandable). The mean (6SEM) understandability rating of the items selected for the reputation condition was 5.3560.09. Fifty-two of the 68 items were presented twice, as they were commonly and independently selected by both groups of evaluators. All subjects were told that they would rate the eight evaluators after the fMRI experiment, and that our aim was to investigate the neural mechanisms underlying first impression formation.
In the experiment, we included conditions in which the subjects viewed the same items but were told that the items represented the impressions of four strangers (two male and two female), in addition to themselves and their partner, in order to investigate intimacy effects on neural activation. At the beginning of the fMRI experiments, the subjects were told that the same eight assessors had evaluated these four strangers. The use of four agents in the stranger condition was similar to a previous study [6]. If one agent was used in the stranger condition, the same groups of adjectives would be used for three persons (self, partner, and stranger). This situation might raise the possibility of subjects noticing the experimenters' manipulation of adjectives. Using four agents in the stranger condition was intended to keep this effect to a minimum. In addition, we confirmed that the subjects did not know the four strangers by interviewing them after the experiments.
To check the validity in terms of impression-formation modulation caused by intimacy, 14  stranger, and fourth stranger, respectively. One-way (target = self/ partner/first stranger/second stranger/third stranger/fourth stranger) repeated measures analysis of variance (rmANOVA) showed a significant main effect (p,0.05). Post-hoc analysis showed no significant differences between pairs formed with each of the four strangers. However, pairs formed of each of the four strangers and self or partner showed significant differences (between self and first stranger or second stranger, p,0.001; partner and fourth stranger, p,0.05; and other pairs, p,0.01). The impression formations of the four strangers were similar and the impression of self or partner was well formed in comparison.
In the fMRI experiments, the subjects were required to evaluate the social desirability of the shown items on a three-point scale (ranging from 3, desirable to 1, undesirable) using the right index, middle, or third fingers during the reputation conditions, irrespective of which person was presented. During the noreputation condition, the subjects were asked to press a button with their right index finger. Subjects were required to respond within 3 s to each stimulus in the reputation and no-reputation conditions. Four seconds after stimulus onset, the next stimulus was presented. The reputation, no-reputation, and rest blocks were each 24 s in length (Fig. 1).
The reputation conditions were repeated four times for the self, partner, and stranger conditions in each fMRI run. The noreputation conditions were repeated two times for the self, partner, and stranger conditions in each fMRI run. In addition, a fixation rest block was repeated two times in each run. There were five runs in each session. The participating couples completed the fMRI session simultaneously using hyper-scanning 3T fMRIs positioned side-by-side. The condition sequence for each run was predetermined and counterbalanced across subjects. Two blocks of self, partner, and stranger conditions were never presented in succession within a run. Prior to the fMRI experiment, the subjects took part in a 2-min practice session, in which no photos were included and the adjectives were different from those used in the fMRI experiment. The total duration of the run was 8 min.

fMRI Data Analysis
We used SPM8 revision 4667 (The Wellcome Trust Centre for Neuroimaging; http://www.fil.ion.ucl.ac.uk/spm) implemented in MATLAB 2010a (MathWorks, Inc., Massachusetts) to analyze the functional images. The first four volumes of each fMRI run were discarded because the MRI signal was unsteady. We initially performed motion correction, normalization to the Montreal Neurological Institute (MNI) template, and spatial smoothing (8 mm). After the realignment processes, we checked the headmovement parameters. The task-related activation was evaluated statistically on a voxel-by-voxel basis using a general linear model at the individual level to generate contrast images, which then were incorporated into random-effects analysis at the group level [18].
In the fMRI data analysis, we excluded data from six subjects; thus, data from 16 subjects were analyzed. To investigate the neural correlates underlying reputation representation, we compared the brain activation between the reputation and noreputation conditions for the self, partner, and stranger conditions. As the mPFC is activated during both positive self reputation [6] and negative self reputation [7], the reputation representation processed in the mPFC should not be influenced by social desirability. As reputation valuation processes were modulated by social desirability [6], we included the average social desirability scores during each reputation-condition block as parametric modulation terms in the individual level analysis for distinguishing reputation representation-related activation from valuation-related activation. The average social desirability score for each block was calculated using the social desirability ratings obtained after the fMRI experiment. After the fMRI experiment, social desirability was evaluated by each subject using a seven-point Likert scale. We then conducted a random-effects analysis using the contrast images produced by the first-level analysis (contrasts of the reputation and no-reputation conditions for the three target types [self, partner, and stranger] and the interaction of the reputation effects).

Behavioral Data Analysis
We conducted statistical analysis of the social desirability ratings and response times acquired during the fMRI experiment. In this analysis, we performed a one-way (target = self/ partner/stranger) rmANOVA of the social desirability scores. We also conducted a two-way (reputation level [reputation/noreputation]6target [self/partner/stranger]) rmANOVA of the response times for the social desirability judgments made during the fMRI experiment.

Rating Results
During the fMRI experiment, the average (6SEM) social desirability ratings during the self, partner, and stranger conditions were 2.6460.05, 2.6960.04, and 2.6460.04, respectively. The one-way (target = self/partner/stranger) rmANOVA did not show a significant main effect of target (p = 0.084) (Fig. 2(A)).
Response times for the social desirability ratings for the reputation condition in the self, partner, and stranger conditions were 1290.7642.6, 1256.2640.3, and 1288.4647.0 ms, respectively. For the no-reputation condition, the response times in the self, partner, and stranger conditions were 600.8629.9, 604.4631.8, and 598.7631.1 ms, respectively. A 2 (reputation level = reputation/no-reputation)63 (target = self/partner/stranger) rmANOVA showed a significant main effect of reputation level (p,0.001) and a significant interaction effect (p,0.05). There was no significant main effect of target (p = 0.415). Post-hoc analyses showed significant differences (p,0.001) for all the reputation level pairs in each target, and between the self-reputation and partnerreputation conditions (p,0.05) (Fig. 2(B)).

fMRI Results
The interaction effects of self vs. stranger and the reputation effect (self (reputation -no-reputation).stranger (reputation -noreputation)) showed significant activations in the mPFC and precuneus in the whole-brain analysis (Table 1 and Fig. 3). The mPFC cluster was located in the dorsal part, and mainly occupied the superior frontal gyrus extending to the frontal pole, suborbital sulcus, and cingulate sulcus. The interaction effects of the partner vs. stranger conditions and the reputation effect (partner (reputation -no-reputation).stranger (reputation -no-reputation)) revealed two significant clusters in the mPFC in the wholebrain analysis (Table 1 and Fig. 4). These two mPFC clusters were located in the dorsal part: the posterior cluster mainly occupied the superior frontal gyrus extending to the cingulate sulcus; the anterior cluster mainly occupied the superior frontal gyrus extending to the frontal pole and suborbital sulcus. The mPFC activation related to the two interaction effects overlapped (Fig. 5). There was no significant activation specific to the strangerreputation effect, and there was no significant difference between the self-and partner-reputation effects.
Scatter diagrams of the average beta values (Fig. 6) did not reveal any effect of handedness on brain activity in the four significant clusters.

Behavioral Data: Social Desirability Ratings
As the rating results were not significantly different between the three conditions (self, partner, and stranger), the social desirability ratings were not influenced by the evaluation target. Thus, the subjects properly rated the social desirability of the presented adjectives. Response times for the social desirability ratings showed that response times during the reputation condition were longer in comparison with the no-reputation condition. The longer response times reflected the increased depth of processing required by the social desirability judgment [6]. These results suggest that subjects accurately performed the task during the fMRI experiment.

Reputation Representation in the mPFC
In the self condition, reputation processing activated the dorsal mPFC in comparison with the stranger condition, consistent with a previous study [6]. This activation was found for the averaged verbal reputation representation of self, and was not modulated by social desirability. Thus, this result supports previously proposed reputation-processing model: the mPFC represents reputation information, which is utilized for valuation processing [8].
The mPFC is a core region for self-referential processing [19,20,21]. Task of self-referential thought is a typical type of selfreflection [22]. In self-reflection processes, subjects evaluate whether presented adjectives conform to self. The dorsal mPFC activation was similar to that in previous studies showing the neural correlates underlying self-reflection [19,20,21,22]. This result confirms that self-reflection processing is necessary for reputation representation.
The mPFC is a key node for social cognition [23]. It represents social learning processes and interacts with social informationsensitive areas such as the anterior cingulate cortex (ACC) [24] and superior temporal sulcus (STS) [25,26]. This kind of learning processes requires theory-of-mind or mentalizing ability [27,28,29,30]. As mPFC activation is related to the mental state as distinct from the physical state [31], the mPFC activation in the present study specifically reflects reasoning while taking into account the perspective of other persons [32] when processing reputation information. In particular, the mPFC represents the accuracy of empathic inference of target emotions [33]; thus, the mPFC activation in the present study might represent perspectivetaking of others, which is required for developing a reputation representation based on evaluations made by others.
A previous study showed that reflected appraisal primarily depends on self-evaluation [11]. This was supported by other studies showing that the neural responses underlying reflected appraisal overlapped with those involved in self-evaluation [34,35]. The mPFC activation in the present study was located near the region where the activation related to reflected appraisal overlapped with the self-evaluation activation in previous studies.   Thus, the mPFC activation might include activation related to other's evaluation in addition to self-evaluation. Filling the gap between others' and one's own perspective (i.e., the anchoring and adjustment process) also activates the mPFC [36]. Therefore, as we predicted, the present results showed that mPFC activation when processing one's own reputation might be related to the comparison between one's own and others' evaluations.

Representation of the Reputation of Intimate Others
In terms of self-and other-reflection, one's own evaluation of a known other (public image) [37,38,39,40] or intimate other (friend or mother) [34,41,42] (other-reflection) in addition to the evaluation of oneself (self-reflection) also activated the mPFC. These results implied that humans form detailed, individualized evaluations of known persons [15] in a way that is similar to selfevaluation. In the present study, compared with strangers, processing of both the verbal reputation of self and the verbal reputation of an intimate other activated the mPFC, which processes reputation representation. This activation pattern is consistent with a previous study, which found common mPFC activation when viewing one's own face and that of an intimate other [43]. As subjects' appraisal of others is individualized depending on the motivation to interact [15], in the present study the visual stimuli in the stranger condition were the only targets for which the subjects did not form their own evaluation. The  common activation in the mPFC during the reputation representation of both oneself and a romantic partner supports our hypothesis that one's own evaluation of a target, which is already formed for intimate others, is necessary for developing a reputation representation for others.
Two clusters in the dorsal mPFC (anterior [BA 10] and posterior [BA 8]) were activated by the reputation effect of partner. The anterior dorsal mPFC provides person perception and mentalizing functions, whereas the posterior dorsal mPFC provides attention-related functions [20]. As reputation represen-tation requires knowledge of how others represent the target person, the anterior dorsal mPFC (with person perception and mentalizing functions) might play an essential role, and could possibly interact with attention-related functions represented in posterior dorsal mPFC.
The overlapping activation in the mPFC for reputations of self and partner was shown by the interaction effects of reputation by target vs. stranger. These overlapping activations were located near the area that processes anchoring and adjustment [36]. Similarity between the perceiver and target, one of the charac- Overlapping activation for the self-and partner-reputation conditions. Significant activation related to the interaction effects of target6reputation in the medial prefrontal cortex (mPFC) is shown. Red and green indicate self-specific (self (reputation -no-reputation).stranger (reputation -no-reputation)) and partner-specific (partner (reputation -no-reputation).stranger (reputation -no-reputation)) interaction effects, respectively. Overlapping activation between these two interaction effects is also shown (yellow). doi:10.1371/journal.pone.0074958.g005 teristics of social closeness, enhances the perceiver's social projection for the target person and weakens stereotype processing [44] through anchoring and adjustment [36]. Reputation representation processes involving anchoring and adjustment require evaluation related to self-and other-referential processing. Overlapping activation was located in the dorsal part of the mPFC, which is responsible for evaluation in self-and otherreferential processing [22]. The present overlapping activation for the reputation of a romantic partner and self indicates that the mPFC processes the reputation of an intimate other by using one's own evaluation of that person and others' evaluations of him/her through anchoring and adjustment similar to reputation processing regarding self.
The mPFC is also known to provide functions for other target person-specific processes, such as trust [25,26]. mPFC activation fluctuates with the amount of trust that one should is statistically predicted to have in others [25,26]. In the present study, the reputation and no-reputation conditions presented images of the same target persons. This might evoke similar trust feelings in subjects. The contrast between the reputation and no-reputation conditions was intended to cancel out such target person-specific processing. This contrast was expected to provide only reputation effects caused by subjects' belief in the adjective being considered by other evaluators. Furthermore, the overlapping activation in the mPFC between the partner and self conditions shows intimacy modulation for reputation representation.

Self-specific Activation
The precuneus was activated specifically during self-reputation processing. The precuneus serves a wide range of functions, including episodic memory retrieval and self-processing operations [45]. As episodic memory is used for the storage and recall of previously experienced events, it has autobiographical reference [46] (i.e., it requires retrieving and constructing information related to the self). Precuneus activity during episodic memory retrieval [47] might reflect such self-related processing. In addition, imagining the future, which also requires elaborating upon information related to oneself, showed overlapping activation with episodic memory retrieval in the precuneus [48]. In terms of self-processing operations, self-referential judgment of adjectives (self-reflection) also activates the precuneus [49]. In addition to these task-dependent activations, the precuneus shows prominent activation during resting states (task-independent activation) [50]. This baseline activation in the precuneus during resting states (default mode-network activation) is associated with the representation of the world around us [50]. The default-mode network includes the mPFC, which was activated during the self and partner reputation conditions in the present study, in addition to the precuneus [50]. These two areas, the mPFC and precuneus, are linked according to diffusion tensor image (DTI) fiber tractography and functional connectivity [51]. In terms of comparisons between self-reflection processes and the defaultmode network, the mPFC shows greater activity for the former, whereas the precuneus extending to the posterior cingulate cortex shows greater activity for the latter and shared activation for self reflection and the default-mode network [52]. This differential activity pattern suggests that the precuneus activity for reputation of self might reflect default network-related activation. Taken together, these findings suggest that the precuneus might play a key role in the mental representation of the self. Thus, in the present study, the precuneus activation that was specific to the selfreputation condition could represent the formation of the mental representation of the self caused by reputation of self.

Conclusion
As predicted, the present results showed that the neural correlates underlying verbal reputation representation of others are modulated by intimacy. This intimacy effect suggests that we represent the reputation of intimate others made by third parties in a way that is similar to how we process reputation of self. Due to the intimate relationship between romantic partners, we develop our own evaluation of intimate others. These results suggest that our own evaluation, together with others' evaluations, is necessary to represent the reputation of others.