Data Sheet 1_Lateralized nigrostriatal dopamine pathway activation promotes early reversal learning.docx

The dorsomedial striatum (DMS) is believed to promote action-outcome associations by integrating cortical and limbic afferents with nigrostriatal dopamine. This functional role for the DMS encompasses reinforcement learning processes traditionally ascribed to the ventral (outcome valuation) and lateral (action-selection) subdivisions of the striatum. Previous studies have shown that DMS dopamine signaling encodes actions and outcomes, often in a lateralized manner (e.g., dopamine release is greater for contralateral actions). To determine how these dynamics evolve with changing action-outcome contingencies, we recorded dopamine axon Ca2+ activity in the DMS during a lateralized reversal learning task in mice. Using a miniaturized fluorescence microscope, we found that dopamine axon Ca2+ activation in the DMS encoded actions and outcomes with a lateral bias, but only very early in reversal learning, immediately after the action-outcome contingency switch. Specifically, we found that dopamine axon activation during contralateral choices and the rewards associated with those choices were only greater than ipsilateral choices and their rewards in the first session of reversal. Over the course of reversal, reward- and choice-evoked dopamine axon activation subsided for contralateral, but not ipsilateral, choices, resulting in no lateral bias after learning the new action-outcome contingency. Consistent with a causal role for these lateralized dynamics, unilateral optogenetic inhibition of the nigrostriatal dopamine pathway impaired contralateral reversal by reducing “win-stay” responses. However, this deficit was transient, occurring only during the first reversal session. Our results suggest that dopamine signaling in the DMS facilitates the exploration of new actions, specifically at the specific moment the previous action-outcome contingency becomes conflicted. In our lateralized reversal learning task, this facilitation was particularly important when the new action and resulting outcome were contralateral to the recorded hemisphere. These findings advance our understanding of how the DMS carries out its ascribed role in action-outcome learning.