Predictive Representations in Hippocampal and Prefrontal Hierarchies.

Area of Science:

• Cognitive Neuroscience
• Computational Neuroscience

Background:

• Human navigation relies on learned relational structures for exploration and goal achievement.
• Existing research on relational knowledge and planning is limited to small-scale, controlled environments.
• Understanding how stored knowledge supports navigation in realistic, continuous settings remains an open question.

Purpose of the Study:

• To investigate how multiscale predictive representations guide naturalistic human navigation.
• To determine the hierarchical organization of these predictive scales in the brain.
• To explore the roles of the prefrontal cortex and hippocampus in large-scale spatial navigation.

Main Methods:

• Model-based representational similarity analysis of neuroimaging data.
• Participants navigated long, realistic paths in a virtual reality environment.
• Analysis focused on pattern similarity to successor points within varying predictive horizons.

Main Results:

• A gradient of predictive horizons was observed across prefrontal and hippocampal regions.
• Anterior prefrontal cortex (PFC) exhibited the largest predictive horizons.
• Posterior hippocampus demonstrated the smallest predictive horizons, with intermediate regions like anterior hippocampus and orbitofrontal cortex falling in between.

Conclusions:

• Multiscale predictive representations are hierarchically organized in the brain, supporting goal-directed navigation.
• Findings provide empirical support for computational models of navigation and hierarchical planning.
• This research advances understanding of cognitive maps in everyday, long-distance navigation.