Characterization of the effect of dopamine on the neural coding of reward-based learning and decision-making

Dopamine has an important role in normal cognition and reward processing, both of which are impaired in disorders involving dopamine dysfunction such as addiction, schizophrenia, and Parkinson’s disease. However, our understanding of the interplay between different aspects of the dopamine system and reward-guided behavior in humans is limited. Food is an important type of reward that is critical for survival and impacts the decisions we make every day. Here, we characterize the relationship between two food-reward related phenotypes and dopamine synthesis capacity (related to tonic dopamine) as well as dopamine D1 and D2 receptor availability in healthy humans. First, we examined the link between dopamine synthesis and receptor availability and body mass regulation in 117 individuals with body mass index (BMI) values ranging from normal to obese. We found that current BMI was related to a pattern of increased dopamine synthesis in the hypothalamus, a region important for homeostatic control of appetite, but decreased dopamine D2 receptor availability in the midbrain, where D2 autoreceptors regulate dopamine release throughout the brain. This suggests that increased BMI is related to a dopamine imbalance between homeostatic drivers of appetite and reward system regulatory control mechanisms that could result in an overactive, unregulated intake of food. Building on this finding, we studied the link between dopamine synthesis capacity and receptor availability and an important food-reward related behavior, foraging. Fifty-seven healthy volunteers completed a computer-based foraging task where we measured their threshold for leaving one group of rewards to search for another in four different reward environments varying from a low to high rate of reward receipt. We found that two particular patterns of dopamine synthesis and receptor availability in the anterior cingulate cortex and basal ganglia were linked to the amount that individuals changed their threshold based on the reward rate of the environment. Finally, since the prefrontal cortex is known to be important for reward-guided behavior, we implemented two methodological advancements aimed to address limitations that make it difficult to measure cortical dopamine in humans with PET imaging. The first method involves partial volume correction and surface-based smoothing in order to increase the signal to noise in the cortex. The second method is a data-driven PET data parcellation and automated reference region selection algorithm to optimize the voxels included in the reference region. In conclusion, we have characterized the dopaminergic contribution of two different foodreward guided phenotypes and have developed two techniques that will aid future research on the role of cortical dopamine. Understanding the neural mechanisms underlying these rewardguided behaviors helps us to not only understand normal behavior, but also serves as a reference for comparison when studying related pathological states.