Rostral Anterior Cingulate Activations inversely relate to Reward Payoff Maximation & predict Depressed Mood

Abstract

Humans make choices based on both reward magnitude and reward frequency. Probabilistic decision making is popularly tested using multi-choice gambling paradigms that require participants to maximize task payoff. However, research shows that performance in such paradigms suffers from individual bias towards the frequency of gains as well as individual differences that mediate reinforcement learning, including attention to stimuli, sensitivity to rewards and risks, learning rate, and exploration vs. exploitation based executive policies. Here, we developed a two-choice reward task, implemented in 186 healthy human subjects across the adult lifespan, to understand the cognitive and neural basis of payoff-based performance. We controlled for individual gain frequency biases using experimental block manipulations and modeled individual differences in reinforcement learning parameters. Simultaneously recorded electroencephalography (EEG)-based cortical activations showed that diminished theta activity in the right rostral anterior cingulate cortex (ACC) as well as diminished beta activity in the right parsorbitalis region of the inferior frontal cortex (IFC) during cumulative reward presentation correspond to better payoff performance. These neural activations further associated with specific symptom self-reports for depression (greater ACC theta) and inattention (greater IFC beta), suggestive of reward processing markers of clinical utility.