Meta-learning is expressed through altered prefrontal cortical dynamics.

Area of Science:

• Neuroscience
• Cognitive Science
• Animal Behavior

Background:

• Survival depends on learning reward availability, which can be complex in natural environments.
• Meta-learning, or learning how to learn, is crucial for adapting to changing reward contingencies.
• Neural mechanisms underlying meta-learning in animals are not well understood.

Purpose of the Study:

• Investigate meta-learning in a spatial foraging task with a novel depletion-repletion rule.
• Examine neural dynamics in the medial prefrontal cortex (mPFC) during meta-learning.
• Understand how mPFC neural activity supports learning new reward strategies.

Main Methods:

• Rats performed a spatial foraging task with a learned depletion-repletion rule.
• Longitudinal, high-density neural recordings were conducted in the medial prefrontal cortex (mPFC).
• Analysis focused on changes in neural dynamics and coding during meta-learning progression.

Main Results:

• Meta-learning induced systematic changes in mPFC neural dynamics, embedding the learned rule.
• Individual mPFC neurons showed mixed coding of task structure and value.
• Population-level dynamics organized into generalizable low-dimensional motifs that were reshaped during meta-learning.
• These dynamics supported both predictive inference and outcome-based value updating.

Conclusions:

• Meta-learning reshapes prefrontal cortex (mPFC) neural dynamics to acquire new reward-learning strategies.
• The mPFC plays a key role in implementing meta-learning through dynamic neural representations.
• These findings advance our understanding of the neural basis of adaptive learning.