Figure 2.

Toward a unifying account of dopamine in value-based choice. (A) Combining the accelerator/brake with the proximity model (Figure 1), we propose that dopamine affects cost-benefit decision making both by enhancing the proximity advantage of proximate low-cost (e.g., sooner rewards) over distant high-cost options (e.g., later rewards) and by implementing a cost control. If no option possesses a proximity advantage over the other (upper panel), our account makes the same predictions as the accelerator/brake account, i.e., increasing both D₁ receptor (D1R) (via enhanced benefit processing) and D₂ receptor (D2R) activity (by increasing the acceptable costs) enhances the preference for high-benefit rewards. If the low-cost option is also more proximate than the high-cost option (lower panel; note that costs and proximity can conceptually be distinguished), low dopamine levels (by reducing the influence of the proximity advantage) and increased D1R activity strengthen the preference for more distant–high benefit rewards. In contrast, high D2R activity enhances both the proximity advantage of the more proximate (low cost) option and the acceptable costs. (B) Within the framework of a drift diffusion–style model of the choice process, D2R activation might increase the proximity advantage of low-cost options by shifting the starting point of the evidence accumulation process toward the boundary of the low-cost option (71). During evidence accumulation, D1R activation strengthens the impact of benefits on the velocity of the accumulation process (drift rate), while higher D2R activation lowers the sensitivity to action costs. Thus, if decision thresholds are low, stronger proximity effects under high D2R activation increase the likelihood of choosing the (proximate) low-cost option, whereas in cases of high decision thresholds, D2R effects on evidence accumulation will result in more choices of the high benefit–high cost option.