Differential reinforcement encoding along the hippocampal long axis helps resolve the explore-exploit dilemma.

Area of Science:

• Neuroscience
• Cognitive Science
• Decision-Making

Background:

• Decisions involve balancing exploiting known options versus exploring new ones.
• Natural environments often feature structured reward distributions, unlike unstructured exploration relying on neocortex, striatum, and amygdala.
• The hippocampus is crucial for creating cognitive maps to navigate and forage in structured environments.

Purpose of the Study:

• To investigate the distinct roles of the posterior hippocampus (PH) and anterior hippocampus (AH) in exploration and exploitation.
• To understand how hippocampal subregions represent information during reinforcement learning with structured rewards.

Main Methods:

• Utilized a reinforcement learning task with a spatially structured reward function in human participants.
• Analyzed brain activity, focusing on differential reinforcement representations in the PH and AH.

Main Results:

• The PH invigorates exploration, while the AH supports the transition to exploitation.
• PH exhibits early, phasic reward prediction error signals.
• AH shows delayed, sustained global value maximum signals and goal cell-like activity.

Conclusions:

• The PH and AH play distinct, complementary roles in navigating structured choice environments.
• Differential neural representations in hippocampal subregions support adaptive exploration-exploitation strategies.
• AH dynamically updates value representations and guides goal-directed navigation.