Year: 2026
Type: article
Abstract: Goal-directed decisions depend on predictions of future outcomes and are thus sensitive to revaluation of expected outcomes, whereas habitual decisions depend on cached action values or stimulusresponse associations and are insensitive to outcome revaluation. Hippocampal neurons encode future states (potentially supporting prediction of future outcomes during goal-directed choice) as well as present states (potentially supporting retrieval of state-dependent action values during habitual choice). Prior studies suggest the role of the hippocampus in goal-directed, outcome-sensitive decision-making, but few studies have examined its role in outcome-insensitive action selection. Here, we tested this using a two-outcome spatial navigation task in which identical turning actions (left or right) led either to preferred milk or nonpreferred water outcomes, depending on the spatial location at which the choice was made. Rats (20 males, 22 females) were trained under conditions that either required hippocampal-dependent discrimination between choice points or allowed hippocampal-independent navigation using visual beacons. Rats then underwent 24 h water restriction, during which separate cohorts underwent Contingency Updating (CU), in which rats re-experienced the maze actionoutcome contingencies while thirsty, or Incentive Learning (IL), in which rats received only a two-bottle preference test while thirsty without re-experiencing maze contingencies. Both cohorts were later tested under thirst on an unbaited maze. CU rats shifted their choices toward water-associated responses, whereas IL rats perseverated on milk-associated responses, even when navigation depended on the hippocampus. These results provide novel evidence that hippocampal-dependent navigational decisions can show insensitivity to outcome revaluation, a widely used behavioral criterion for habitual control.
Source: Journal of Neuroscience
Authors: Shiyun Wang, Ryan Grgurich, Hugh T. Blair
Institution: University of California, Los Angeles
Language: English
Topic: Memory and Neural Mechanisms
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: hybrid
OpenAlex ID: W7214111655