Striatal cell-type-specific encoding stability and reorganization for goal-directed behavior and habit

Goal-directed decision making relies on learned action-outcome associations to allow actions to adapt when outcomes change. With repetition, behavioral control shifts toward habit, enabling more automatic action execution at the expense of flexibility. Much is unknown about how the brain learns and supports goal-directed behavior and what changes as habits form. Here, we longitudinally tracked and manipulated the activity of two major dorsomedial striatum populations, D1+ and D2/A2A+ neurons, to reveal their activity and function during goal-directed learning and habit formation. Across learning, DMS D1+ neurons stably encode instrumental actions, carry information about behavioral engagement and organization, and develop and maintain action-reward representational similarity and conjunctive activation. This persists even after behavioral control transitions to habit. In contrast, one ensemble of DMS A2A+ neurons transiently encodes actions during early learning and another slowly starts to encode actions as habits form. Both DMS D1+ and A2A+ neuronal activity mediates the learning processes underlying goal-directed control, but only DMS D1+ neurons maintain support of adaptive, goal-directed decision making. These data reveal a cell-type-specific dissociation between stable v. reorganizing striatal action representations and, independently, persistent v. transient contributions to goal-directed control.