Learning and Representation of Hierarchical Concepts in Hippocampus and Prefrontal Cortex

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Human cognition flexibly uses conceptual knowledge at varying abstraction levels, suggesting hierarchical organization.
• The neural mechanisms for acquiring and representing this hierarchical structure remain largely unknown.
• Previous research focused on category knowledge representation, not its acquisition process.

Purpose of the Study:

• To investigate the computations underlying the acquisition and representation of hierarchical conceptual knowledge.
• To explore the roles of the hippocampal-prefrontal system in learning hierarchical structures.
• To differentiate learning computations and representation in key brain regions.

Main Methods:

• Utilized a novel tree-like categorization task with 32 human participants.
• Employed computational model comparisons to assess hierarchical learning.
• Combined behavioral data with functional magnetic resonance imaging (fMRI).

Main Results:

• Medial prefrontal cortex (mPFC) tracks the accumulation of hierarchical conceptual knowledge.
• mPFC and hippocampus support trial-to-trial updating of concepts.
• Connectivity changes revealed rostro-lateral PFC represents the final hierarchical structure.

Conclusions:

• mPFC and hippocampus integrate evidence accumulation and updates for hierarchical concept formation.
• Rostro-lateral prefrontal cortex ultimately represents the learned hierarchical conceptual knowledge.
• This study elucidates the neural basis of hierarchical concept acquisition and representation.