Passively learned spatial navigation cues evoke reinforcement learning reward signals.

Area of Science:

• Neuroscience
• Cognitive Psychology
• Computational Neuroscience

Background:

• Theories of navigation suggest humans and rodents form cognitive maps, but whether this involves reinforcement learning remains debated.
• Allocentric navigation, or the creation of cognitive maps, is a key aspect of spatial learning.
• Understanding the learning mechanisms behind cognitive maps is crucial for advancing spatial cognition research.

Purpose of the Study:

• To investigate the role of reinforcement learning in allocentric navigation.
• To determine if event-related brain potentials (ERPs) can reveal reinforcement learning during cognitive map acquisition.
• To examine the neural correlates of value assignment to navigational cues.

Main Methods:

• Participants navigated a virtual environment using allocentric, egocentric-response, or egocentric-cue strategies.
• After navigation, participants viewed cue images associated with each strategy.
• Event-related brain potentials (ERPs), specifically reward positivity, were measured in response to these cues.

Main Results:

• A reward positivity was elicited when allocentric navigators viewed cues previously associated with the goal location.
• This suggests that cues predicting the goal in allocentric navigation acquire long-term value.
• The findings link reinforcement learning mechanisms to the formation of cognitive maps.

Conclusions:

• Reinforcement learning contributes to the acquisition of allocentric navigation and cognitive map generation.
• Navigational cues can retain predictive value after the learning phase, as indicated by ERPs.
• This study provides neural evidence supporting the integration of reinforcement learning in spatial cognition.