PROJECT SUMMARY Dysregulated activity within the neural networks that underlie reward and executive functions can increase vulnerability to compulsive drug use. Repetitive transcranial magnetic stimulation (rTMS) can selectively modulate neural activity within both the reward and the executive control networks. Empirical evidence indicates that both inhibitory rTMS to the ventromedial prefrontal cortex (VMPFC, a key node in the reward circuits) and excitatory rTMS to the dorsolateral prefrontal cortex (DLPFC, a key node of the executive control network) can be therapeutic in individuals with SUDs. However, responses to both treatments are highly variable and rTMS’ therapeutic utility for SUDs remains limited. Personalizing rTMS interventions targeting neuromarkers of dysregulated activity within the reward and cognitive control networks is likely to reduce treatment response variability and improve treatment outcomes. We have recently identified two robust neuromarkers of reward responses to drug-related cues and cognitive control during drug-related decisions that reliably predict nicotine self-administration. The goal of this application is to determine the extent to which these neuromarkers moderate responses to rTMS. Our central hypothesis is that smokers with high reactivity to drug-related cues will be more likely to reduce nicotine self-administration after inhibitory rTMS to the VMPFC, whereas smokers with low cognitive control during drug-related decisions will be more likely to reduce nicotine self-administration after excitatory rTMS to the DLPFC. In line with the scope of PAR-19-282 (Exploratory Clinical Neuroscience Research on SUDs), we will use a phased research approach to test our hypothesis. In the R61 phase, we will determine the extent to which the reward reactivity neuromarker (Aim 1) and the cognitive control (Aim 2) neuromarker moderate rTMS treatment effects. To test our milestones (i.e., observing a medium or higher effect size in at least one of the two aims), we will use Bayesian statistical methods. Using a Bayesian approach will allow us to obtain calibrated probabilities of treatment outcomes and make a principled go/no-go decision in moving forward to the R33 phase. In the R33 phase, we will determine the synergistic effect of the two neuromarkers in moderating rTMS treatment effects (Aim 3) and we will create a Bayesian classifier to personalize future rTMS interventions for SUDs (Aim 4). We anticipate that upon its successful conclusion, this Phase II trial will contribute to illuminate the psychophysiological mechanisms that drive compulsive drug use and will yield the fundamental knowledge needed to efficiently develop new personalized rTMS interventions for SUDs and other disorders characterized by poor impulse control.