Many neuropsychiatric disorders involve compulsive behavior, including obsessive-compulsive behavior, obesity, eating disorders, alcoholism, and addiction. Compulsive behavior appears to arise from impaired top- down control of striatal learning mechanisms, particularly involving circuits between the orbitofrontal cortex (OFC) and the caudate nucleus (CN). Theories of decision-making have emphasized a complex relationship between fast, automatic, habitual processes and slower, deliberative, goal-directed choices. These distinct processes are thought to map on to distinct frontostriatal circuits with automatic processes and goal-directed choices linked to different striatal regions. Theoretical accounts of OFC control have argued that it is responsible for providing state information to CN, ensuring that the valuation of reward outcomes is contextually appropriate. Recent psychophysics results have shown that habitual responses are prepared simultaneous with goal-directed responses, but are inhibited to allow the slower, goal-directed response to occur. This raises the possibility that OFC may have a gating function, similar to that which has been posited for more dorsolateral frontal regions, whereby it inhibits habitual responses to allow goal-directed behaviors. In the current grant, we will test the hypothesis that OFC is responsible for inhibiting habitual responses in the striatum when top-down control and more deliberative decision-making is required. We will use a novel behavioral task that enables us to manipulate the amount of top-down control required for a decision using two conflicting reward contingencies. This is manifest as an increase in the amount of time necessary to make the decision as well as the number of saccades that the animals make to either option. In addition, we have spent the last year developing methods to lower the new primate Neuropixels probes into deep targets within the brain, including OFC and CN, and for managing the large amounts of data that these probes generate allowing us to use population-level decoding to investigate the dynamics of these cognitive processes with high temporal and single-trial resolution. In addition, we will test a second hypothesis that OFC accomplishes striatal inhibition via coherence in the alpha band. We have previously shown, via closed-loop microstimulation, that reward-based learning depends on theta coherence between OFC and the hippocampus. However, recent work has emphasized the role of the alpha band in mediating inhibitory processes, both in the frontal cortex and in posterior sensory cortex. We will use our expertise with closed-loop microstimulation to test whether OFC inhibits CN via alpha coherence during top-down control. Understanding the dynamics of OFC-CN interactions will lay the groundwork for building devices that can meaningfully interact with these circuits.