Interoceptive, Affective, and Cognitive Control Networks that Determine Self-Regulation and Reinforcement Learning of Smoking Decisions - PROJECT SUMMARY Although quitting smoking offers significant health benefits for all age groups1, smoking cessation is notoriously difficult to achieve2. Smokers commonly report that the greatest challenge in quitting is inhibiting and managing the urge to smoke (i.e., self-regulation), which is often triggered by internal (interoception) or external (exteroception) cues associated with nicotine (a primary reinforcer) through reinforcement learning3. While self- regulation and reinforcement learning (RL) processes are crucial cognitive mechanisms that contribute to successful smoking cessation outcomes, previous behavioral and neuroimaging studies have not provided specific information about the computational and neurobiological bases of these key decision-making processes that could be used to prevent self-control failures (e.g., a single puff after abstinence) at a particular moment (e.g., a stressful situation) or condition accompanied by previously associated hyper cue-reactivity. Using computational model-based multi-modal (EEG, fMRI) neuroimaging and novel custom-made EEG/MRI- compatible e-cigarette smoking devices, we will investigate the neurocomputational mechanisms that determine self-regulation and RL in smoking decisions at both short-term (trial-by-trial) and long-term (over one year) levels. One hundred eighty e-cigarette or other electronic nicotine delivery systems (ENDS) users (early adulthood; 21-35 years) who have thought about quitting smoking at least once in the past year will complete two sessions. In the EEG session, participants, having abstained from smoking overnight, will make ‘real’ smoking choices about whether or not to take a puff of an e-cigarette under three conditions (emotional distress, cognitive overload, and no stress). We expect that brain network interactions among interoception (insula), affective/motivational (ventral striatum, vmPFC), and cognitive control (dlPFC) systems will predict trial-by-trial self-control smoking decisions. In the fMRI session, participants will complete probabilistic RL and extinction tasks using ‘real’ e-cigarette smoking and money as rewards. We hypothesize that brain responses in the interoceptive (insula), affective/motivational (ventral striatum, vmPFC), and cognitive control (dlPFC) regions will explain the dysregulation of RL and extinction with smoking rewards compared to money rewards. In the one-year follow-up, we will systematically test whether participants’ smoking habit changes can be predicted by our computational model parameters and EEG/fMRI susceptibility and resilience measures. The knowledge gained from our study, which (1) predicts real smoking regulation decisions and (2) explains the learning and extinction processes of cue-reactivity at both short-term and long-term levels, will have strong ecological validity and provide valuable, transformative insights for developing novel interventions that may prevent smoking lapses before they occur, as often imagined in science fiction. Beyond smoking cessation treatments, our project will also enhance scientific understanding of other self-control and hyper cue-reactivity- related maladaptive lifestyle behaviors (e.g., obesity, alcohol or drug abuse) that increase the risk of cancer.
