Temporally and Functionally Distinct Contributions to Value Based Choice Along the Anterior-Posterior Dorsomedial Striatal Axis

While the dorsoventral and mediolateral organization of striatum has resolved clear functional distinctions, far less is known about how the anterior-posterior striatal axis contributes to behavioral control. We explore this within the dorsomedial striatum (DMS), a key region for value-based choice, by comparing population neuronal activity and function within anterior (A-DMS) and posterior (P-DMS) subregions while mice operantly seek reward. Neural recordings show that P-DMS encoded action values and strategy information prior to choice selection while A-DMS activity represented recently selected choices and their anticipated values via a dynamic population reorganization immediately following action selection. Optogenetic perturbations were consistent with these temporally distinct coding properties as unilateral manipulation of the P-DMS prior to choice biased choice contralaterally in a value-dependent manner and unilateral inhibition of the A-DMS following choice impaired future value-based action selection. Using anterograde tracing, we found that the A-DMS and P-DMS projected to a common region within the ventromedial substantia nigra pars reticulata (vmSNr), which contained value-related signals combining aspects of upstream DMS processing. Together, our results support a model for temporally distributed influence on value-based choice across the anterior-posterior axis of the DMS.