PDF(2194 KB)
PDF(2194 KB)
PDF(2194 KB)
面孔可信度和社会价值取向对信任与信任学习的共同作用与机制*
The Joint Effect and Mechanism of Facial Trustworthiness and Social Value Orientation on Trust and Trust Learning
现有信任研究多聚焦于探究面孔可信度或社会价值取向对信任的独立影响，缺乏对二者在单次信任与信任学习中共同作用的系统探讨，且在使用计算模型方法探究信任学习的动态机制时未能充分考虑社会价值取向的作用。研究结合信任博弈范式与强化学习模型，以回应上述研究缺口。研究1发现，面孔可信度与社会价值取向分别正向影响单次信任，二者不存在共同作用，互惠预期在二者的正向影响中起中介作用。研究2发现，社会价值取向显著调节初始面孔可信度与回馈概率对信任学习的共同作用。强化学习模型成功捕获了以往研究发现的关键计算机制特征，但不同社会价值取向个体的模型参数无显著差异，表明信任学习存在尚未揭示的计算机制。研究发展了整合外部线索与内在特质的信任研究框架，为经典信任理论提供了实证依据，并从计算模型视角拓展了信任学习中个体差异的研究路径。
Current research on trust has primarily examined the independent effects of facial trustworthiness and social value orientation (SVO), without systematically investigating their joint influence on both one-shot trust and trust learning, and their underlying mechanisms. Furthermore, existing studies employing computational models to explore the dynamic processes of trust learning have not adequately incorporated SVO as a key factor of individual difference. Grounded in the person-context tnteraction theory and the dyadic model of trust development, facial trustworthiness as an external cue and SVO as an intrinsic trait should jointly shape trust behavior. Based on classical trust theories, such as the social norm theory and moral norm theory, reciprocity expectation may serve as the underlying psychological mechanism through which both factors influence trust. Similarly, drawing on the neuropsychological economic model of trust, the dynamic interaction between initial facial trustworthiness and feedback probability in trust learning is likely to vary as a function of an individual’s SVO, and such differences should be reflected in the computational modeling of trust learning. To address these research gaps and test these theoretical hypotheses, the present study combined the trust game paradigm with reinforcement learning modeling across two experiments to systematically examine the joint effects of facial trustworthiness and SVO at different stages of trust development (one-shot trust and trust learning) and to quantify key computational mechanisms underlying trust learning.
Experiment 1 employed a one-shot trust game to examine whether facial trustworthiness and SVO interactively influence one-shot trust and whether reciprocity expectation mediates their effects. A 2 (facial trustworthiness: high vs. low) × 2 (SVO: prosocial vs. proself) mixed design was adopted, with facial trustworthiness as a within-subjects factor and SVO as a between-subjects factor. The dependent variables were the amount invested and reciprocity expectation. The results showed that both facial trustworthiness and SVO significantly and positively predicted one-shot trust, but their interaction was not significant. Reciprocity expectation mediated the effects of both factors on trust.
Experiment 2 used a repeated trust game to examine whether the dynamic interaction between initial facial trustworthiness and feedback probability during trust learning is moderated by SVO. A 2 (SVO: prosocial vs. proself) × 2 (initial facial trustworthiness: low vs. high) × 2 (feedback probability: high vs. low) mixed design was implemented, with the first two factors as within-subjects variables forming consistent and inconsistent conditions, and SVO as a between-subjects variable. Twelve candidate reinforcement learning models were constructed and compared to identify the best-fitting model. Statistical analyses of model parameters were conducted to elucidate the specific mechanisms by which SVO influences trust learning. The results revealed that SVO significantly moderated the interactive effects of initial facial trustworthiness and feedback probability: proself individuals trusted low-trustworthiness trustees more under high feedback probability but trusted high-trustworthiness trustees more under low feedback probability. In contrast, prosocial individuals consistently trusted high-trustworthiness trustees more regardless of feedback probability. Model comparisons showed that Model 4c, which incorporated three key mechanisms identified in prior research, provided the best fit. However, SVO had no significant effect on learning rates, the facial trustworthiness influence coefficient, or the decay rate.
The findings demonstrate that facial trustworthiness and SVO exert independent effects on one-shot trust but exhibit a joint influence in trust learning, where the dynamic interaction between initial facial trustworthiness and feedback probability is moderated by SVO. This suggests that their synergistic effect emerges specifically in dynamic interactive contexts. The study verifies that reciprocity expectation serves as a common psychological mechanism through which both factors influence trust, providing empirical support for classical trust theories. The reinforcement learning model successfully captured key computational characteristics of trust learning from previous studies. Individuals adjusted learning rates based on the consistency between initial facial trustworthiness and feedback probability. Initial facial trustworthiness exerted a persistent influence on subsequent learning. Also, the impact of prediction errors from behavioral feedback on value expectations diminished over time. This study incorporated SVO into a computational modeling framework for the first time. It also revealed that trust learning may involve additional computational processes beyond the known mechanisms. SVO might influence learning through these processes. This warrants further investigation in future research.
信任 / 信任学习 / 面孔可信度 / 社会价值取向 / 强化学习模型
trust / trust learning / facial trustworthiness / social value orientation / reinforcement learning model
| [1] |
陈瀛, 徐敏霞, 汪新建. (2020). 信任的认知神经网络模型. 心理科学进展, 28(5), 800-809.
信任是指一方在基于对另一方行为期望的基础上愿意冒一定的风险, 以期在将来得到积极结果的心理过程。近年, 认知神经取向的研究对信任行为引起的特定脑区激活进行了考察, 却忽略了大规模脑网络在信任过程中的整体作用。在总结前人研究的基础上提出信任的认知神经网络模型, 并从认知神经网络视角对信任行为进行解释和整合。在模型中, 信任行为是动力系统、情感系统和认知系统相互作用的结果, 并分别与奖励网络、显著网络、中央执行网络和默认网络等神经网络激活有关。此外, 模型还强调信任行为的反馈机制, 以此构成完整的建构模型。模型阐明了心理系统与中枢神经网络之间的对应关系, 从认知神经角度解释了信任行为发生的心理机制和神经基础。
|
| [2] |
高青林, 周媛. (2021). 计算模型视角下信任形成的心理和神经机制-基于信任博弈中投资者的角度. 心理科学进展, 29(1), 178-189.
人际信任渗透在社会交互的各个方面, 是促进和维持合作的重要基石。以往研究者借助信任博弈范式, 主要探讨了人际信任的理论模型、生物基础和影响因素等方面。近年来, 研究者开始将计算模型应用于信任博弈的数据分析中, 深入挖掘人际信任行为背后的心理机制, 将计算模型与神经影像技术结合, 加深对信任行为背后脑机制的理解。目前将计算模型应用于信任博弈范式中的研究主要针对“信任是如何形成的”这一科学问题, 未来要进一步发展计算模型方法, 结合非侵入性脑刺激技术, 应用于精神疾病人群中, 以深入理解正常和异常信任形成的心理和神经机制。
|
| [3] |
刘长江, 郝芳. (2011). 不对称社会困境中社会价值取向对合作的影响. 心理学报, 43(4), 432-441.
|
| [4] |
谈昊天, 李泽清, 伍珍. (2025). AI与人谁更可信? 基于强化学习建模的人智重复信任博弈. 心理科学, 48(4), 920-932
|
| [5] |
王才康. (2000). 基于项目的方差分析探讨. 心理学报, 32(2), 224-228.
|
| [6] |
王阳, 温忠麟. (2018). 基于两水平被试内设计的中介效应分析方法. 心理科学, 41(5), 1233-1239.
|
| [7] |
|
| [8] |
|
| [9] |
|
| [10] |
Recent efforts to understand the mechanisms underlying human cooperation have focused on the notion of trust, with research illustrating that both initial impressions and previous interactions impact the amount of trust people place in a partner. Less is known, however, about how these two types of information interact in iterated exchanges. The present study examined how implicit initial trustworthiness information interacts with experienced trustworthiness in a repeated Trust Game. Consistent with our hypotheses, these two factors reliably influence behavior both independently and synergistically, in terms of how much money players were willing to entrust to their partner and also in their post-game subjective ratings of trustworthiness. To further understand this interaction, we used Reinforcement Learning models to test several distinct processing hypotheses. These results suggest that trustworthiness is a belief about probability of reciprocation based initially on implicit judgments, and then dynamically updated based on experiences. This study provides a novel quantitative framework to conceptualize the notion of trustworthiness.Copyright © 2010 Elsevier Inc. All rights reserved.
|
| [11] |
|
| [12] |
Trust is essential for a secure and flourishing social life, but many economic and philosophical approaches argue that rational people should never extend it, in particular to strangers they will never encounter again. Emerging data on the trust game, a laboratory economic exchange, suggests that people trust strangers excessively (i.e., far more than their tolerance for risk and cynical views of their peers should allow). What produces this puzzling "excess" of trust? We argue that people trust due to a norm mandating that they show respect for the other person's character, presuming the other person has sufficient integrity and goodwill even if they do not believe it privately. Six studies provided converging evidence that decisions to trust follow the logic of norms. Trusting others is what people think they should do, and the emotions associated with fulfilling a social duty or responsibility (e.g., guilt, anxiety) account for at least a significant proportion of the excessive trust observed. Regarding the specific norm in play, trust rates collapse when respect for the other person's character is eliminated as an issue.
|
| [13] |
Facial appearances can powerfully influence adults' trust behaviour, despite limited evidence that these cues constitute honest signals of trustworthiness. It is not clear, however, whether the same is also true for children. The current study investigated whether, like adults, 5-year-olds and 10-year-olds are more likely to place their trust in partners that look trustworthy than those that look untrustworthy. A second, closely related question was whether children also explicitly value the information from face cues when making trust decisions. We investigated these questions using Token Quest: an economic trust game that gave participants the opportunity to make investments with a series of partners who might (or might not) repay their trust with large returns. These interactions occurred under different conditions, including one in which participants were shown the face of each partner and another in which they could 'purchase' access to faces with a portion of their investment capital. Results indicated that, like adults, 10-year-old children selectively placed their trust in those partners they perceived as looking trustworthy and many were willing to 'pay' to purchase access to these face cues during the trust game. We observed a similar profile of trust behaviour in 5-year-olds, with no significant group difference in the impact of face cues on behaviour across the three age groups. Together, these findings indicate that the influence of face cues on trust behaviour emerges early, and highlight a capacity for sophisticated social cognition in young children. © 2014 John Wiley & Sons Ltd.
|
| [14] |
|
| [15] |
|
| [16] |
|
| [17] |
|
| [18] |
|
| [19] |
|
| [20] |
Trust pervades nearly every social aspect of our daily lives, and its disruption is a significant factor in mental illness. Research in the field of neuroeconomics has gained a deeper understanding of the neuropsychoeconomic (NPE) underpinnings of trust by combining complementary methodologies from neuroscience, psychology, and economics. However, a coherent model of trust that integrates separate findings under a conceptual framework is still lacking. Here, we sketch out an integrative NPE model that explains how the interactions of psychoeconomic components engage domain-general large-scale brain networks in shaping trust behavior over time. We also point out caveats of current research approaches and outline open questions that can help guide future transdisciplinary investigations for a better understanding of the neuropsychology of trust.Copyright © 2018 Elsevier Ltd. All rights reserved.
|
| [21] |
|
| [22] |
|
| [23] |
Facial trustworthiness can influence trust behavior. The outcome may be concordant (a person with a trustworthy face who engages in reciprocal behavior) or conflicting (a person with a trustworthy face who does not reciprocate). This study investigated the temporal features of processing different outcomes following trustworthy or untrustworthy faces. Thirty-one college students participated in a decision-making task in which they were shown a face and were asked to judge whether the person would behave in a trustworthy or an untrustworthy manner in a hypothetical scenario, and feedback was then provided about the target person's behavior in the scenario, while the event-related potential of the electroencephalography of participants' brains was recorded. In the feedback presentation stage, there was a significant interaction between face type and feedback. Trustworthy faces with losses induced the largest feedback-related negativity (FRN), whereas trustworthy faces with gains induced the smallest FRN. For untrustworthy faces, there was no significant difference between losses and gains. Trustworthy faces with losses and untrustworthy faces with gains induced more positive P300. FRN is possibly sensitive to both negative outcomes and expectation violations. P300 is mainly sensitive to expectation violations. Individuals have high positive expectations for trustworthy faces. When the behavior of a trustee with trustworthy face was negative, so as to violate the trustor's expectation, it will elicit a more serious conflict in the mind of the trustor.
|
| [24] |
Previous event-related potential (ERP) studies employing monetary gambling tasks have demonstrated that the brain responds differentially not only to one's own gain and loss but also to the others' gambling outcomes. Empathy and motivational significance are implicated in the processes of outcome evaluation. This study is to explore to what extent the brain activity is modulated by the interpersonal relationship between the individual and the other agent, who can be a friend or a stranger. Brain potentials were recorded while the participant observed reward feedback to his/her own, his/her friend's, or a stranger's performance in a gambling task. The magnitude and latency of the effect on an early ERP component, the FRN, did not differ between the friend- and the stranger-observation conditions, whereas a late component, the P300, was modulated not only by reward valence but also by the interpersonal relationship between the observer and the other agent. These findings suggest that brain responses in outcome evaluation may be divided into an earlier semi-automatic process and a later cognitive appraisal process and that the interpersonal relationship comes into play mostly in the late attention-sensitive stage.2009 Elsevier Ltd. All rights reserved.
|
| [25] |
Humans exhibit considerable variation in how they value their own interest relative to the interests of others. Deciphering the neural codes representing potential rewards for self and others is crucial for understanding social decision-making. Here we integrate computational modeling with functional magnetic resonance imaging to investigate the neural representation of social value and the modulation by oxytocin, a nine-amino acid neuropeptide, in participants evaluating monetary allocations to self and other (self-other allocations). We found that an individual's preferred self-other allocation serves as a reference point for computing the value of potential self-other allocations. In more prosocial participants, amygdala activity encoded a social-value-distance signal; that is, the value dissimilarity between potential and preferred allocations. Intranasal oxytocin administration amplified this amygdala representation and increased prosocial behavior in more individualistic participants but not in more prosocial ones. Our results reveal a neurocomputational mechanism underlying social-value representations and suggest that oxytocin may promote prosociality by modulating social-value representations in the amygdala.
|
| [26] |
|
| [27] |
Recently, cross-cultural facial-expression recognition has become a research hotspot, and a standardised facial-expression material system can significantly help researchers compare and demonstrate the results of other studies. We developed a facial-expression database of Chinese Han, Hui and Tibetan ethnicities. In this study, six basic human facial expressions (and one neutral expression) were collected from 200 Han, 220 Hui and 210 Tibetan participants who lived in these regions. Four experts on each ethnicity evaluated the facial-expression images according to the expressions, and only those achieving inter-rater agreement were retained. Subsequently, 240 raters evaluated these images according to the seven emotions and rated the intensity of the expressions. Consequently, 2980 images were included in the database, including 930 images of Han individuals, 962 images of Hui individuals and 1088 images of Tibetan individuals. In conclusion, the facial-expression database of Chinese Han, Hui and Tibetan people was representative and reliable with a recognition rate of over 60%, making it well-suited for cross-cultural research on emotions.© 2019 International Union of Psychological Science.
|
| [28] |
|
| [29] |
|
| [30] |
We used event-related potentials (ERPs) to tap the temporal dynamics of first impressions based on face appearance. Participants were asked to evaluate briefly presented faces for trustworthiness and political choice. Behaviorally, participants were better at discriminating faces that were pre-rated as untrustworthy. The ERP results showed that the P100 component was enhanced for untrustworthy faces, consistently with the view that signals of potential threat are given precedence in neural processing. The enhanced ERP responses to untrustworthy faces persisted throughout the processing sequence and the amplitude of early posterior negativity (EPN), and subsequent late positive potential (LPP) was increased with respect to trustworthy faces which, in contrast, elicited an enhanced positivity around 150 ms on frontal sites. These ERP patterns were found specifically for the trustworthiness evaluation and not for the political decision task. Political decision yielded an increase in the N170 amplitude, reflecting a more demanding and taxing structural encoding. Similar ERP responses, as previously reported in the literature for facial expressions processing, were found throughout the entire time course specifically elicited by faces explicitly judged as untrustworthy. One possibility might be that evolution has provided the brain with a 'special toolkit' for trust evaluation that is fast and triggers ERPs related to emotional processing.
|
| [31] |
|
| [32] |
What motivates people when they make decisions and how those motivations are potentially entangled with concerns for others are central topics for the social, cognitive, and behavioral sciences. According to the postulate of narrow self-interest, decision makers have the goal of maximizing personal payoffs and are wholly indifferent to the consequences for others. The postulate of narrow self-interest-which has been influential in economics, psychology, and sociology-is precise and powerful but is often simply wrong. Its inadequacy is well known and efforts have been made to develop reliable and valid measurement methods to quantify the more nuanced social preferences that people really have. In this paper, we report on the emergence and development of the predominant conceptualization of social preferences in psychology: social value orientation (SVO). Second, we discuss the relationship between measurement and theory development of the SVO construct. We then provide an overview of the literature regarding measurement methods that have been used to assess individual variations in social preferences. We conclude with a comparative evaluation of the various measures and provide suggestions regarding the measures' constructive use in building psychologically realistic theories of people's social preferences.
|
| [33] |
Learning the value of stimuli and actions from others - social learning - adaptively contributes to individual survival and plays a key role in cultural evolution. We review research across species targeting the neural and computational systems of social learning in both the aversive and appetitive domains. Social learning generally follows the same principles as self-experienced value-based learning, including computations of prediction errors and is implemented in brain circuits activated across task domains together with regions processing social information. We integrate neural and computational perspectives of social learning with an understanding of behaviour of varying complexity, from basic threat avoidance to complex social learning strategies and cultural phenomena.
|
| [34] |
|
| [35] |
|
| [36] |
|
| [37] |
|
| [38] |
Trust is a key component of social interaction. Older adults, however, often exhibit excessive trust relative to younger adults. One explanation is that older adults may learn to trust differently than younger adults. Here, we examine how younger (N = 33) and older adults (N = 30) learn to trust over time. Participants completed a classic iterative trust game with 3 partners. Younger and older adults shared similar amounts but differed in how they shared money. Compared to younger adults, older adults invested more with untrustworthy partners and less with trustworthy partners. As a group, older adults displayed less learning than younger adults. However, computational modeling suggests that this is not because older adults learn differently from positive and negative feedback than younger adults. Model-based fMRI analyses revealed several age- and learning-related differences in neural processing. Specifically, we found that older learners (N = 19), relative to older non-learners (N = 11), had greater reputation-related activity in metalizing/memory areas while making their decisions. Collectively, these findings suggest that older adult learners use social cues differently from non-learners.Copyright © 2023 The Author(s). Published by Elsevier Inc. All rights reserved.
|
| [39] |
|
| [40] |
|
| [41] |
Decades of research document individual differences in prosocial behavior using controlled experiments that model social interactions in situations of interdependence. However, theoretical and empirical integration of the vast literature on the predictive validity of personality traits to account for these individual differences is missing. Here, we present a theoretical framework that identifies 4 broad situational affordances across interdependent situations (i.e., exploitation, reciprocity, temporal conflict, and dependence under uncertainty) and more specific subaffordances within certain types of interdependent situations (e.g., possibility to increase equality in outcomes) that can determine when, which, and how personality traits should be expressed in prosocial behavior. To test this framework, we meta-analyzed 770 studies reporting on 3,523 effects of 8 broad and 43 narrow personality traits on prosocial behavior in interdependent situations modeled in 6 commonly studied economic games (Dictator Game, Ultimatum Game, Trust Game, Prisoner's Dilemma, Public Goods Game, and Commons Dilemma). Overall, meta-analytic correlations ranged between -.18 ≤ ρ̂ ≤.26, and most traits yielding a significant relation to prosocial behavior had conceptual links to the affordances provided in interdependent situations, most prominently the possibility for exploitation. Moreover, for several traits, correlations within games followed the predicted pattern derived from a theoretical analysis of affordances. On the level of traits, we found that narrow and broad traits alike can account for prosocial behavior, informing the bandwidth-fidelity problem. In sum, the meta-analysis provides a theoretical foundation that can guide future research on prosocial behavior and advance our understanding of individual differences in human prosociality. (PsycINFO Database Record (c) 2019 APA, all rights reserved).
|
| [42] |
|
| [43] |
The authors adopt an interdependence analysis of social value orientation, proposing that prosocial, individualistic, and competitive orientations are (a) partially rooted in different patterns of social interaction as experienced during the periods spanning early childhood to young adulthood and (b) further shaped by different patterns of social interaction as experienced during early adulthood, middle adulthood, and old age. Congruent with this analysis, results revealed that relative to individualists and competitors, prosocial individuals exhibited greater levels of secure attachment (Studies 1 and 2) and reported having more siblings, especially sisters (Study 3). Finally, the prevalence of prosocials increased--and the prevalence of individualists and competitors decreased--from early adulthood to middle adulthood and old age (Study 4).
|
| [44] |
|
| [45] |
Mimicking real world situations, the Chicken Game allows scientists to examine human decision-making when the outcome is not entirely within one person's control. In this social dilemma task, two players independently choose either to safely cooperate with, or riskily aggress against, the other player, and the unique combination of their choices specifies the outcome for each. Coupling the Chicken Game with psychophysiological measures, we confirmed our two hypotheses: that an individual perceives an outcome as most negative when she chooses to cooperate and the other player violates that trust and aggresses, and that motivational salience of an outcome is greater when an individual chooses to aggress and when she gains money. Collectively, the data demonstrate the utility of pairing true social dilemma tasks like the Chicken Game with psychophysiological measures to better understand decision-making.Copyright © 2013 Elsevier B.V. All rights reserved.
|
| [46] |
|
| [47] |
The recent years have witnessed a dramatic increase in the use of reinforcement learning (RL) models in social, cognitive and affective neuroscience. This approach, in combination with neuroimaging techniques such as functional magnetic resonance imaging, enables quantitative investigations into latent mechanistic processes. However, increased use of relatively complex computational approaches has led to potential misconceptions and imprecise interpretations. Here, we present a comprehensive framework for the examination of (social) decision-making with the simple Rescorla-Wagner RL model. We discuss common pitfalls in its application and provide practical suggestions. First, with simulation, we unpack the functional role of the learning rate and pinpoint what could easily go wrong when interpreting differences in the learning rate. Then, we discuss the inevitable collinearity between outcome and prediction error in RL models and provide suggestions of how to justify whether the observed neural activation is related to the prediction error rather than outcome valence. Finally, we suggest posterior predictive check is a crucial step after model comparison, and we articulate employing hierarchical modeling for parameter estimation. We aim to provide simple and scalable explanations and practical guidelines for employing RL models to assist both beginners and advanced users in better implementing and interpreting their model-based analyses.© The Author(s) 2020. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.
|
/
| 〈 |
|
〉 |
