dACC stimulation stabilizes behavior without affecting reward processing

Balancing exploration and exploitation is crucial for navigating unpredictable, ever-changing environments. Exploitation stabilizes decision-making to maximize rewards while exploration introduces variability, destabilizing behavior to search for new opportunities. Previous studies have implicated the dorsal anterior cingulate cortex (dACC) in exploratory decision-making, but also in tracking important decision-making variables like reward. Critically, the relationship between dACC and these variables remains ambiguous because neither link has been tested causally. Here, we stimulated dACC during multiple tasks that allowed us to differentiate the effects of dACC microstimulation on exploration and reward processing. First, two rhesus macaques performed a decision-making task---a multi-arm bandit---where they selected one of three targets by making a saccade towards it. Although the targets were visually identical, each was associated with a dynamic reward probability. This encouraged monkeys to use reward information to choose between exploiting good targets and exploring alternatives that could offer greater long-term gains. We then stimulated dACC and a control region, dorsolateral prefrontal cortex (dlPFC), during inter-trial intervals on random trials. Stimulating dACC but not dlPFC stabilized the behavior by reducing the likelihood of starting to explore. This was not due to a change in reward processing: stimulation stabilized behavior regardless of reward history. To assess whether dACC stimulation stabilized behavior because it was itself hedonically rewarding, the monkeys completed another task, choosing between reward cues that were either paired or unpaired with dACC stimulation. dACC stimulation did not systematically influence subjective target value. These findings causally and selectively implicate dACC stimulation in regulating exploration.