Abstract
When navigating through space, the brain must integrate sensory information with past experiences to choose behaviors that are most likely to produce a positive outcome. Striatal medium spiny neurons (MSNs) expressing dopamine receptor 1 (D1) and dopamine receptor 2 (D2) receive sensory and motor information from cortical and midbrain regions to modulate locomotion; how these cells are differentially activated in changing environments can contribute to sensory-dependent behavior selection. Dopamine release is evoked during unpredicted reward events, and training with a conditioned stimulus shifts the dopamine response from the reward to the reward-predictive stimulus. Models of dopamine dependent reinforcement learning postulate that dopamine can serve as a ‘teaching’ signal which can enable sensory stimuli to become associated with reward by strengthening and weakening synaptic connections onto MSNs. Modulation of dopamine release during reinforcement learning has differential long-term effects on D1 and D2 MSN activity patterns in the striatum which could be the basis for Go and No-go decision making in various environments. Sensory stimuli that have consistently led to a positive outcome in the past would lead to strong long-term potentiation in D1 or long-term depression in D2 MSNs receiving repeated coincident glutamate and dopamine input, leading over time to repetition of movements leading to reward. Using fiber photometry, the activity pattern of D2 MSNs during virtual navigation was found to be different in familiar versus novel environment, while the D1 population maintained consistent firing patterns across sensory contexts.