人类非常热衷于了解自身在不同方面是如何被评价的。无论是赞扬的评论还是批评的评论，个人都会很热衷于去获得。例如，我们会频繁的查看社交网络平台（如微信、脸书、推特等）去看有多少人点赞了自己发表的内容，以及大家是怎么评论的。并且，网络平台中人们同样乐于参与不同种类的“性格测评”，甚至愿意付出一定程度的金钱代价来了解自己是如何被评价的。然而，个体这种愿意付出代价来获知评价的行为是如何产生的，以及其中的计算与分子机制仍然不清楚。因此，本研究首先假设获知行为具有主观价值（Subjective value），并结合计算建模，通过2个探索性行为实验（实验1与实验2）一个大样本重复实验（实验3），检验并验证价值化社会性评价与非社会性评价的认知过程与计算机制。进而，为了探索其分子机制，本研究结合一个元分析实验（实验4）与两个行为药理实验（实验5与实验6）揭示了社会认知相关的神经多肽——催产素在价值化评价中的作用。

为了探索这一问题，本研究基于以往研究发展了Pay to know这一范式。这一范式中，个体需要选择是否付出一定的金钱来获得外界的评价。具体而言，被试需要先在第一天参加第一阶段，即提供个人的照片以及完成一份个人信息相关的问卷，并被告知相关性信息给予另一个人来进行评价（实验1与实验3的社会性评价条件），或提供给电脑进行分析与评价（实验2与实验3的非社会性评价条件）。然后，被试需要参加第二阶段的行为实验，完成Pay-to-know实验任务。每个试次中，被试会看到一个特质性的词语（积极特质或消极特质），并选择是否获得评价，不同的选项对应不同的金钱数量。我们采取心理物理法与计算建模的方法进行分析，通过对行为数据拟合无知规避参数揭示其中的计算过程，并且通过对行为与反应时数据进行漂移模型的拟合揭示信息累计的过程。

两个行为实验（实验1&2）的结果显示获知社会性评价与非社会性评价均具有主观价值。对于非社会性评价，积极与消极评价具有同等价值。并且，计算建模的结果显示两种价值共享无知规避的参数，并且在反应时内信息累计的速率一致。相反，对于社会性评价，积极与消极评价具有分离的无知规避参数，而信息累计的过程中积极评价的累计速率快于消极评价，说明了二者分离的计算过程。这也使得，个体赋予积极社会性评价的价值高于消极社会性评价。

元分析研究（实验4）揭示了，鼻喷催产素在社会与情绪认知过程中促进了双侧杏仁核、尾状核、颞上回的活动，同时降低了双侧杏仁核的活动。本研究进一步发现了鼻喷催产素对于健康群体神经活动的影响受到了效价的调节作用，即催产素促进了积极加工过程中尾状核的活动，而降低了消极加工过程中杏仁核的活动。通过对比健康人群与患者人群，结果显示鼻喷催产素均降低了两个人群中双侧杏仁核的活动，但是对患者人群的影响更强。在增强的效应中，催产素促进了健康人群中双侧杏仁核、尾状核与颞上回的活动，但是促进了患者人群中背侧扣带回的活动。结合鼻喷催产素的药理调控实验（实验5&6）中，结果显示催产素的效应同样受到的评价效价的调节，即鼻喷催产素降低了个体对于消极社会性评价的价值以及无知规避的程度，而对积极社会性评价的影响受到了抑郁特质的调节。同时非社会性反馈的价值没有受到催产素的影响。

本研究的研究结果促进了对获知评价对不同动机的认识。具体而言，获知社会性评价中表现的积极偏差可能由获得社会赞许、社会支持的动机驱使；而获知非社会性评价时表现的无偏差，可能说明获知行为由获得准确、客观的自我图像驱使。元分析结果揭示了效价对于催产素效应的重要作用，以及催产素对于社会认知障碍患者人群的神经活动的作用存在特异性。进一步，本研究结果显示效价对于社会性评价的调节作用在鼻喷催产素条件下进一步提升，即催产素降低了个体赋予消极社会性评价的价值，提升了高抑郁特质个体赋予积极社会性评价的价值。这个结果揭示了催产素促进社会适应的作用，降低了消极社会信号的价值，并恢复了积极社会信号对于高抑郁特质的价值，暗示了催产素的临床价值。

Humans are curious about how the self is evaluated, either praised or criticized, even at a personal cost. For example, people spend over 70% conservation time in expressing their options and knowing others’ options on the self. We frequently check social networking sites (such as Facebook, Twitter, etc.) to see how many people ‘like’ their pages or how others comment on their recent posts, and are willing to take online personality quiz to know how artificial intelligence evaluates on us. However, it remains unclear what drives such costly-to-know human behavior. In this research, with discovery and replication samples in a series of studies, we propose and provide cognitive and computational supports for the hypothesis that knowing evaluations on the self is valuable to individual and such subjective value placed on evaluation and release of unknown aversion drive the costly-to-know behavior. We also investigated molecular substrates underlying the valuation process, by conducting a meta-analysis study and two pharmacological studies to examine the role of social neuropeptide – oxytocin.

Here, we develop a novel pay-to-know choice task where participants trade off monetary reward against the opportunity of knowing. Specifically, participants were first invited to a preparation session to complete self-introduction as materials for social (i.e., other people) and non-social (i.e., a computer program) evaluations. Then, participants came to the main experiment session and completed the pay-to-know choice task. For each trial of the pay-to-know task, participants were first presented with a single aspect (a positive or negative trait word) on which he could receive an evaluation and then had to decide whether they wanted to know the evaluation of the presented aspect or not. The two options were associated with different monetary rewards. We then performed psychophysical analysis, computational modeling, and drift diffusion modeling, to quantify the subjective value individuals placed on opportunity for knowing evaluation and investigate the computations and evidence accumulating process.

The results from 3 studies lent cognitive and computational support for our hypothesis. Specifically, participants equally valued positive and negative non-social evaluations. However, individuals computed independent unknown aversion towards positive and negative social evaluations, showed higher drift rate basis for positive than negative evaluation during evidence accumulating process, and placed a higher value on the opportunity to know another person’s evaluation on positive than negative aspects.

The meta-analysis, using Activation Likelihood Estimation, on 66 intranasal-oxytocin functional-MRI studies showed that intranasal oxytocin increased neural activity in bilateral amygdala, left caudate head, and superior temporal gyrus (STG) in healthy individuals and increased dorsal anterior cingulate (dACC) activity in patients. Moreover, oxytocin decreased amygdala activity in both patients and healthy volunteers, but to a greater degree in patients than healthy volunteers. Furthermore, the oxytocin effect in healthy volunteers was modulated by valence. We showed that intranasal oxytocin increased parahippocampus activity and decreased amygdala activity during negative social and affective processing. During positive socio-affective processes, brain activity in caudate head was increased by intranasal oxytocin. Then, the pharmacological studies showed that the valence-dependent valuation of the social evaluation was facilitated by oxytocin, a neuropeptide linked to social feedback learning and valuation processes, which decreased the value of negative social evaluation. Moreover, individuals scoring high in depression trait undervalued positive social evaluation, which is normalized by oxytocin.

We revealed the psychological and computational processes underlying self-image formation and updating. Specifically, we uncovered different driven forces behind to know social and non-social evaluations. When facing non-social evaluation, individuals are motivated to acquire an accurate self-image via knowing how the self is objectively evaluated whereas social evaluation may guide us to approach social approval and avoid potential social rejection. Our meta-analysis study provides convergent neural substrates for intranasal oxytocin effect and neuropsychological mechanism underlying oxytocin effect on social and affective processes. The common and different effects of oxytocin on patients and healthy volunteers and the modulation of oxytocin effects by valence have critical implications. Combining oxytocin administration, the pharmacological studies also revealed oxytocin as a key molecule for the value computation of social (rather than non-social) evaluation. The modulatory effect of depression on oxytocin suggests a role of oxytocin in normalizing hypo-valuation of positive social feedback in depression.