Effects of Caffeine on Reinforcement Learning in Healthy Adults Using PET/MRI

Overview

This research study aims to determine whether and how caffeine intake affects learning process through reward feedback compared to placebo. The data acquired from this study would improve our understanding on the consequence and mechanism of caffeine intake in the aspect of learning process. Participants will perform a reinforcement learning task (i.e. Probabilistic Selection Task) and a motor inhibition task (i.e. Go/NoGo task) in a brain scan. The scan will be done with the Siemens Biograph mMR positron emission tomography (PET)/ magnetic resonance imaging (MRI) 3 Tesla scanner. The PET allows us to see the changes in the "reward signals" - dopamine - in the brain using a radioactive dye called \[11C\]Raclopride. The MRI, on the other hand, enables us to take detailed pictures of the brain activities during cognitive tasks using a high-powered magnet. Reviewing these pictures will help us understand the influence of caffeine on reward signals and brain activities during the learning process.

Adenosine 2A receptors (A2AR) colocalize with and exerts allosteric antagonism to dopamine D2 receptors (D2R) by co-forming functional heterodimers in the striatum. Preliminary studies using positron emission tomography (PET) with \[11C\]Raclopride have shown increased D2/D3R availability by A2AR antagonism with caffeine and decreased D2/D3R availability by enhanced adenosine signaling during sleep deprivation, supporting the notion of A2AR-D2R interactions in vivo. However, how A2AR-D2R interactions contribute to D2R-mediated neurocognitive functions is scarcely investigated. Reinforcement learning, a dopamine-mediated cognitive process crucially involved in various human behaviors including habit, preference, belief, and resistance to change, is often found altered in dopamine-associated disorders. For instance, hyper-dopaminergic function in the striatum, as observed in psychosis, leads to a reduction in reward learning and a blunted task-related neural activity. Through the antagonistic effect of A2AR on D2R signaling, blocking A2AR can potentially enhance D2R-mediated negative reinforcement, a.k.a. a "no-go" response. In rodents, A2AR agonists diminish reinforcement of psychostimulants, while an A2AR antagonist can facilitate reward-seeking effects of reinforcers. Hence, in this double-blind randomized crossover study, the investigators aim to use caffeine, an adenosine antagonist as well as a commonly used psychostimulant by nearly 80% of the worldwide population, to examine whether blocking A2AR will enhance D2R-modulated reinforcement learning/no-go responses through modulating D2R signaling pharmacologically. The long-term goal of this study is to further the understanding of molecular mechanisms related to A2AR-D2R heterodimers and the clinical potential of modulating A2AR-D2R interactions. Twelve young healthy non-smokers will enroll in this study. Each participant will undergo a caffeine and a placebo condition. In each condition, participants will first go through a 6-day ambulatory washout period where participants will be asked to abstain from caffeinated dietary, alcohol, and drugs, and stay in regular bed- and wakeup time. On day 7, a PET/fMRI scan will take place at noon, and a caffeine or placebo tablet will be administered orally 20 mins prior to the scan. Simultaneous PET/fMRI will be used to examine the association between the neurochemical changes (i.e., D2/D3R availability as quantified by \[11C\]Raclopride) and the hemodynamic responses (i.e., task-related blood oxygen level-dependent fMRI activity) during reinforcement learning in the caffeine condition compared to placebo. It is hypothesized that enhanced D2/D3R availability mediates the facilitating effect of caffeine on reinforcement learning. Specifically, the investigators expect that caffeine will enhance fMRI responses in reward-related brain regions, and that the increased fMRI response will positively correlate with a change in D2/D3R availability.