PROJECT SUMMARY/ABSTRACT Imbalance between goal-directed (GD) and habitual (H) behavior plays a central role in the etiology of substance use disorder (SUD)1,2. These behaviors are regulated by distinct frontostriatal regions3,4: orbitofrontal cortex (OFC)5, ventromedial prefrontal cortex (vmPFC)6, dorsolateral prefrontal cortex (dlPFC)7, and the rostral medial striatum (rMS) for GD behavior;; premotor cortex (PM)6, supplementary motor area (SMA)8, and the caudal lateral striatum (cLS) for H behavior. Furthermore, recent work demonstrates a central role for the ventrolateral prefrontal cortex (vlPFC) in arbitrating whether behavior is under the control of the GD or H system in a given situation and facilitating transitions between the two behaviors9,10. Elucidating the anatomy underlying this circuitry is critical to furthering insights about the neural basis of maladaptive habit formation in addiction. The first goal of these studies is to identify the corticostriatal anatomical substrates that are in a position to a) integrate the multiple functional inputs needed for GD and H behavior, respectively, and b) facilitate transitions between the two behaviors. The second goal of these studies is to examine how these substrates may be aberrant in SUD patients. Aim 1 will use anterograde anatomical tract-tracing in non-human primates (NHPs) to identify candidate substrates, by locating areas of the striatum where the cortical projections for each behavior converge;; Aim 2 will use retrograde anatomical tract-tracing in NHPs to corroborate the frontostriatal connections found in Aim 1 and will quantify their strength. Aim 3 will then a) use anatomy-guided functional connectivity MRI (fcMRI) to localize these anatomical substrates in humans and b) assess how functional connectivity amongst these key anatomical substrates is unbalanced in SUD patients. Recent work from our lab has identified loci in the striatum that receive disproportionately numerous inputs from functionally diverse prefrontal regions11,12, referred to as hubs. Hubs are areas of unusually high connectivity to other brain regions and are thought to be central for integrating and distributing information across multiple, diverse regions13,14. We hypothesize that H1. OFC, vmPFC, dlPFC and vlPFC projections form a hub in the rMS, and that PM, SMA and vlPFC projections form a hub in the cLS, representing potential anatomical substrates to integrate the inputs needed for GD and H behavior, respectively. As a corollary, we hypothesize that H2. the vlPFC has projections to both the GD and H striatal hubs, representing anatomical substrates that could allow vlPFC arbitration of transitions between the two behaviors. Finally, we hypothesize that H3. SUD patients - who exhibit an overreliance on H behavior2,15 - will have an abnormal ratio of vlPFC-rMS ...