Extended Data Fig. 10: Model of Goal-Selective Disinhibition to Promote Learning of Reward Locations.

This simplified schematic shows how PV interneurons (blue) and pyramidal (red) cells’ firing changes around reward locations as animals learn reward locations (pink). Our data is consistent with a model in which coordinated decrease in inhibition around reward locations develop rapidly to promote excitatory representations and reactivation of reward locations. Top, When a naïve animal approaches and traverse a new reward location (track schematic above shows animals location in the track), NS interneurons inhibit pyramidal cells with no clear reliable response to the reward location (“No experience” row). After an animal has received a few rewards, NS interneurons exhibit coordinated decreases in firing around the reward locations (“Received some rewards” row). These goal-selective decreases co-occur with and are required for the refinement and stabilization of excitatory place fields around goal locations. Over seconds NS interneuron activity decreases before increases in pyramidal activity. Over trials and days, NS interneurons and pyramidal cells response around reward locations develop together. Bottom, once animals have learned the reward locations, NS interneuron firing not only decreases in reward locations but also before reward locations. Such reward-predictive decreases in inhibition are well-suited to facilitate learning about locations that lead to reward.