Social equality in the number of choice options is represented in the ventromedial prefrontal cortex

Abstract

A distinct aspect of the sense of fairness in humans is that we care not only about equality in material rewards but also about equality in nonmaterial values. One such value is the opportunity to choose freely among many options, often regarded as a fundamental right to economic freedom. In modern developed societies, equal opportunities in work, living, and lifestyle are enforced by antidiscrimination laws. Despite the widespread endorsement of equal opportunity, no studies have explored how people assign value to it. We used functional magnetic resonance imaging to identify the neural substrates for subjective valuation of equality in choice opportunity. Participants performed a two-person choice task in which the number of choices available was varied across trials independently of choice outcomes. By using this procedure, we manipulated the degree of equality in choice opportunity between players and dissociated it from the value of reward outcomes and their equality. We found that activation in the ventromedial prefrontal cortex (vmPFC) tracked the degree to which the number of options between the two players was equal. In contrast, activation in the ventral striatum tracked the number of options available to participants themselves but not the equality between players. Our results demonstrate that the vmPFC, a key brain region previously implicated in the processing of social values, is also involved in valuation of equality in choice opportunity between individuals. These findings may provide valuable insight into the human ability to value equal opportunity, a characteristic long emphasized in politics, economics, and philosophy.

Keywords: neuroeconomics; reward system; social comparison.

Figure 1.

Two-person choice task. A, A trial started with the presentation of an “opportunity stimulus” that indicated the number of choice options available to a participant (SELF) and confederate (OTHER). During the subsequent decision period, each player chose one of the options available to them (the confederate's choice was actually determined by computer). The reward outcomes of their choices were indicated by cards (a face card was associated with monetary rewards whereas a deuce was associated with no reward). In the example shown in the figure, there are two options available to the participant, and four to the confederate. Upon choice, the participant's chosen option is highlighted, followed by the revealed choice of the confederate, followed by simultaneous showing of both of their outcomes (“no reward” for the participant and “reward” for the confederate in this case). B, Opportunity-stimulus conditions. The numbers of choice options for SELF and OTHER were manipulated independently. C, The regression model used to analyze the neural responses to the opportunity stimuli. The 3 × 3 matrices represent the nine opportunity-stimulus conditions shown in B. The same model was also used to analyze subjective emotional pleasantness (Fig. 2B). The β coefficients indicate how the dependent variables (R_{(s, o)}: the neural responses or subjective emotional pleasantness) of an individual are sensitive to (1) the number of available options for the participant (β_{OP_SELF}), (2) those for the confederate (β_{OP_OTHER}), and (3) the “choice equality” or the degree of equality in the numbers of available options between the two players (β_{OP_EQUAL}), quantified using the absolute value of their difference.

Figure 2.

Sensitivity of subjective emotional pleasantness on the numbers of options available. A, Self-reported emotional pleasantness of opportunity stimuli. Participants rated the subjective emotional pleasantness of each of the nine combinations of the opportunity stimuli immediately after the scan. B, The effects on the subjective emotional pleasantness of the number of options for SELF (β_{OP_SELF}), that for OTHER (β_{OP_OTHER}), and the equality in option numbers (i.e., choice equality) between the two players (β_{OP_EQUAL}, quantified using the absolute value of the difference between self and other options). Error bars depict SEM; *p < 0.05 (one-sample t test, two-tailed).

Figure 3.

Neural responses in the VS. A, The VS as identified by the independent reward-localizer task. The VS was significantly activated (p < 0.05, FWE-corrected within the VS anatomical mask) during the MID task in response to potential monetary gains. Image is shown at a stringent threshold (p < 0.0001, whole-brain FWE-corrected; k ≥ 10) for display. B, Activation in the VS in response to the opportunity stimuli. Neural β values were averaged within a 6 mm radius sphere surrounding the peak voxel determined by the group-level random-effects analysis for the reward-localizer task. Error bars depict SEM; *p < 0.05 (one-sample t test, two-tailed). C, D, Relations between neural responses and subjective emotional pleasantness sensitivities. The neural β for the number of options for SELF minus that for OTHER (β_{OP_SELF} − β_{OP_OTHER}) was positively correlated with that derived from the subjective emotional pleasantness across participants (C). On the other hand, the neural β for choice equality (β_{OP_EQUAL}) was not correlated with that derived from the subjective emotional pleasantness (D). Pearson correlation coefficients (r) and associated p values (two-tailed) are shown in the figure.

Figure 4.

Neural responses in the vmPFC. A, The vmPFC as identified by the independent reward-localizer task. The vmPFC was significantly activated (p < 0.05, FWE-corrected within the vmPFC anatomical mask) in response to monetary gain versus no-gain outcomes during the MID task. Image is thresholded at p < 0.001, uncorrected; k ≥ 10 for display. B, Activation in the vmPFC in response to the opportunity stimuli. Neural β values were averaged within a 6-mm-radius sphere surrounding the peak voxel determined by group-level random-effects analysis for the reward-localizer task. Error bars depict SEM; *p < 0.05 (one-sample t test, two-tailed). C, D, Relations between neural responses and subjective emotional pleasantness sensitivities. The neural β for the number of options for SELF minus that for OTHER (β_{OP_SELF} − β_{OP_OTHER}) was not significantly correlated with that derived from the subjective emotional pleasantness across participants (C). On the other hand, the neural β for choice equality (β_{OP_EQUAL}) was positively correlated with that derived from subjective emotional pleasantness (D). Pearson correlation coefficients (r) and associated p values (two-tailed) are shown in the figure.