Learning to represent time in striatal circuits

About This Grant

Project Summary Learning to predict the time of salient environmental stimuli, such as rewards, is essential for survival. During classical conditioning, animals learn when to expect a rewarding stimulus (e.g., food) following a predictive cue (e.g., odor). The striatum is implicated in associative reward learning, movement, and timing. Furthermore, the predominant striatal neuron types, medium spiny projection neurons (MSNs) expressing dopamine receptor type 1 (D1) or 2 (D2), both contribute to these diverse behavioral functions. However, the specific roles of D1 and D2 MSNs in learning to represent time are unknown. The research goal of my training grant is to elucidate how temporal coding by ventral striatal D1 and D2 MSNs evolves during learning and the extent to which these distinct cell types contribute to the learning of timed motor behavior. My central hypothesis is that (1) learning improves D1 and D2 MSN temporal coding in response to a reward-predictive cue, and (2) D1 and D2 MSNs exhibit opposing, complementary roles for learning to initiate precisely timed movements. To test this hypothesis, I will utilize a classical conditioning behavioral paradigm in which mice learn to lick in anticipation following a reward- predictive cue. In either the early or late stages of learning, I will use in vivo electrophysiological recordings from optogenetically identified D1 and D2 MSNs, behavioral assessment of the effect of perturbing these cell types, and computational methods to decode time from neural population activity. In Aim 1, I will record D1 and D2 MSN dynamics in different stages of learning and decode time using a machine learning-based pattern classifier. Aim 1 analyses will make comparisons of temporal decoding performance between D1 and D2 MSNs, early and late stages of learning, and distinct temporal intervals. In Aim 2, I will elucidate the causal roles of D1 and D2 MSN activity for learning to produce precisely timed, reward-conditioned licking movements. I will optogenetically inhibit the activity of D1 or D2 MSNs throughout classical conditioning, and I will determine the extent to which these perturbations impact the execution of conditioned licking behavior (e.g., lick initiation time and lick probability). The expected outcomes are that perturbing D1 and D2 MSN activity throughout learning will impair both the temporal precision and mean time of cue-evoked lick initiation. Specifically, I expect that there will be differential effects of D1 and D2 MSN perturbations, such that inhibition of D1 and D2 MSNs increases and decreases mean lick onset time, respectively. These outcomes would suggest that a precise balance of D1 and D2 MSN activity underlies behavioral timing during associative learning. Overall, this project will provide insights on the temporal nature of how the brain processes reward during experience-dependent learning.