Differential effects of positive versus negative drug reinforcement on contextual coding

Re-exposure to drug-associated spatial contexts often provokes drug craving and relapse. The hippocampus plays a critical role in drug associative learning, with a recent study reporting that drug-context associations were encoded in a specific group of CA1 place cells1. In addition to position (P) coding, both head direction (H) and speed (S) coding in the hippocampus2,3 were shown for encoding a given context. These navigational variables can be encoded conjunctively in place cells to support the acquisition of context-specific memories3,4. Contexts associated with either drug rewards or withdrawal can lead to drug-seeking behavior. However, it is unknown how these opposing drug effects change the contextual coding of the hippocampus through multiple navigational variables. To address this question, we performed morphine (MO) conditioned place preference (CPP) and conditioned place aversion (CPA) to model the contextual association with drug reward and withdrawal, meanwhile imaged the Ca2+ activity of individual CA1 neurons using miniscopes. To dissociate neural coding from correlated navigational variables, we used a linear-non-linear Poisson (LN) model5 to account for deceptive correlations and unbiasedly identify neurons that encode single vs. mixed navigational variables. These variables include position (P), head direction (H), speed (S) and all their possible combinations (mixed selectivity, i.e., PH, PS, HS, and PHS). As a result, we found drug rewards (CPP) primarily decreased the number of position coding only place cells, leaving mixed selective place cells unaffected, whereas drug withdrawal (CPA) primarily increased the number of mixed selective place cells but not position coding only place cells. Neither effect was seen in a sucrose CPP experiment, despite the same behavioral change. Together, using an LN model, our work revealed the distinct impact of opposing drug reinforcement on the contextual coding in the hippocampus, providing insights into how hippocampal computations support the transition from voluntary to compulsive drug use.