A comparison of reward-contingent neuronal activity in monkey orbitofrontal cortex and ventral striatum: guiding...

We have investigated how neuronal activity in the orbitofrontal-ventral striatal circuit is related to reward-directed behavior by comparing activity in these two regions during a visually guided reward schedule task. When a set of visual cues provides information about reward contingency, that is, about whether or not a trial will be rewarded, significant subpopulations of neurons in both orbitofrontal cortex and ventral striatum encode this information. Orbitofrontal and ventral striatal neurons also differentiate between rewarding and non-rewarding trial outcomes, whether or not those outcomes were predicted. The size of the neuronal subpopulation encoding reward contingency is twice as large in orbitofrontal cortex (50% of neurons) as in ventral striatum (26%). Reward-contingency-dependent activity also appears earlier during a trial in orbitofrontal cortex than in ventral striatum. The peak reward-contingency representation in orbitofrontal cortex (31% of neurons), occurs during the wait period, a period of high anticipation prior to any action. The peak ventral striatal representation of reward contingency (18%) occurs during the go period, a time of action. We speculate that signals from orbitofrontal cortex bias ventral striatal activity, and that a flow of reward-contingency information from orbitofrontal cortex to ventral striatum serves to guide actions toward rewards.