Social network modulation of reward-related signals

Abstract

Everyday goals and experiences are often shared with others who may hold different places within our social networks. We investigated whether the experience of sharing a reward differs with respect to social network. Twenty human participants played a card guessing game for shared monetary outcomes with three partners: a computer, a confederate (out of network), and a friend (in network). Participants subjectively rated the experience of sharing a reward more positively with their friends than the other partners. Neuroimaging results support participants' subjective reports, as ventral striatal BOLD responses were more robust when sharing monetary gains with a friend as compared to the confederate or computer, suggesting a higher value for sharing with an in-network partner. Interestingly, ratings of social closeness covaried with this activity, resulting in a significant partner × closeness interaction; exploratory analysis showed that only participants reporting higher levels of closeness demonstrated partner-related differences in striatal BOLD response. These results suggest that reward valuation in social contexts is sensitive to distinctions of social network, such that sharing positive experiences with in-network others may carry higher value.

Figure 1.

Task structure. MRI participants played a card guessing task (Delgado et al., 2000) with one of three partners on a given trial—a computer, a confederate, or a close friend—as indicated by a picture at the top of the screen. MRI participants' roles alternated across runs between making the guesses (player run) and watching their partners make the guesses (observer run). A green check mark indicated a correct guess and a monetary gain of $4.00; a red “X” indicated an incorrect guess and a monetary loss of $2.00. Importantly, all outcomes were to be shared equally between MRI participants and their partners.

Figure 2.

Post-session ratings. MRI participants completed subjective ratings of their experiences during the task. Ratings of excitement for winning with each partner demonstrated a main effect of partner: participants were significantly more excited to earn shared positive outcomes with their friend as compared to the confederate or computer (±SEM; **p < 0.016, *p < 0.025).

Figure 3.

Neuroimaging results: outcome phase. a, A contrast of positive > negative outcomes revealed bilateral ventral striatal activity (q(FDR) < 0.01). b, A 2 (role) × 3 (partner) repeated-measures ANOVA on extracted mean parameter estimates from right ventral striatum in response to positive outcomes demonstrated a significant main effect of partner, showing strongest BOLD activation when sharing positive outcomes with a friend (±SEM; **p < 0.016, *p < 0.025). (c). Including MRI participants' IOS ratings of their friends as a covariate with right ventral striatal BOLD activation to positive outcomes revealed a significant partner × IOS interaction: only high closeness (n = 13; orange bars) participants show significant differences in right ventral striatal BOLD activation as a function of partner (±SEM; **p < 0.016, *p < 0.05). Low closeness participants (n = 7; gray bars) do not exhibit these differences, though they tend to show increased BOLD responses regardless of partner.

Figure 4.

Neuroimaging results: response phase. a, A 2 (role) × 3 (partner) ANOVA during the response phase revealed a main effect of partner (q(FDR) < 0.05) in medial prefrontal cortex. A, Anterior; P, posterior. b, Extracted mean parameter estimates reveal that this effect was driven by significantly increased activity during the response phase of friend trials, as compared to confederate or computer trials (±SEM; **p < 0.016, *p < 0.025).