Topic: reward revaluation

Navigating rodents show “forward sweeps” - sequential reactivation of non-local neural representations of future states starting from their current position. While direct evidence remains scant, it is widely thought that such forward sweeps support model-based (MB) evaluation of actions. However, a key challenge in testing this hypothesis is the absence of experimental paradigms that allow for precise measurement of MB action evaluation. Notably, typical assessments of MB evaluation, i.e. flexibility following changes to rewards (reward revaluation), conflate reliance on full MB evaluation wit

A variety of algorithms have been proposed to account for goal-directed planning, operationalized as sensitivity to reward revaluation i.e. flexibility in the face of reward changes. These algorithms use different representations which trade off computational cost with flexibility in the face of changes in the world. In particular, model-based (MB) reinforcement learning uses tree-search through 1-step relational representations that are often thought of as ‘semantic’. In contrast, models like the successor representation (SR) store representations of sequential experience. These algorithms ca