Computational Substrates of Social Value in Interpersonal Collaboration

Abstract

Decisions to engage in collaborative interactions require enduring considerable risk, yet provide the foundation for building and maintaining relationships. Here, we investigate the mechanisms underlying this process and test a computational model of social value to predict collaborative decision making. Twenty-six participants played an iterated trust game and chose to invest more frequently with their friends compared with a confederate or computer despite equal reinforcement rates. This behavior was predicted by our model, which posits that people receive a social value reward signal from reciprocation of collaborative decisions conditional on the closeness of the relationship. This social value signal was associated with increased activity in the ventral striatum and medial prefrontal cortex, which significantly predicted the reward parameters from the social value model. Therefore, we demonstrate that the computation of social value drives collaborative behavior in repeated interactions and provide a mechanistic account of reward circuit function instantiating this process.

How learning shapes the empathic brain

“…Unexpected positive outcomes resulting from help of another person elicit prediction error signals, that is, signals that resemble the ones known to drive reinforcement learning in the monkey brain (17,18). Our results indicate that basic learning mechanisms are also used during complex social learning, which is in line with previous studies (13,(35)(36)(37)(38). Going further, our findings show how classical learning mechanisms shape otherregarding motivational states such as empathy.…”

“…Forty healthy men (mean age, 22.7; SE, 0.41) participated in the study. They were randomly assigned to the experimental and the control group with no age difference between the groups [t (38) …”

Why we stay with our social partners: Neural mechanisms of stay/leave decision-making

“…Moreover, it has been demonstrated that social context can modulate neural reward signals (i.e., Bault, Pelloux, Fahrenfourt, Ridderinckhof, & van Winden, 2015;Fareri, Chang, & Delgado, 2015;Fareri & Delgado, 2014;Fareri, Niznikiewicz et al, 2012;Inagaki et al, 2015) and neural representations of decision value (Strombach et al, 2015). Interpretations of these findings rely thus far on a theorized psychological preference for social connection (e.g., need to belong; Baumeister & Leary, 1995;Bowlby, 1969).…”

Combating Fake News: An Investigation of Information Verification Behaviors on Social Networking Sites

“…Dissimilarity between individuals negatively influences assessments of entitativity, the extent to which the parties perceive that they are part of a single group [8], and is associated with lower levels of trust [53]. Research suggests the relationship between collaboration as a member of a perceived group and increased levels of trust is deeply rooted in our neurobiology [14], and this relationship is well-supported in the extant literature from a wide variety of academic disciplines [33]. Thus, we posit:…”

“…This is because unlike the more superficial approaches to discerning similarity, the identification of common belief structures requires knowledgebased familiarity, a particularly strong antecedent of trust [19]. Because SNSs are used to project a desired social identity that exemplifies the beliefs of the individual [14], posts serve as cues which allow network members to evaluate the degree of cognitive homogeneity that exists between themselves and other network members [20]. Individuals that perceive themselves as having similar opinions and thoughts as other network members are likely to connect more deeply than individuals that perceive themselves as different [48].…”