Two Anatomically and Computationally Distinct Learning Signals Predict Changes to Stimulus-Outcome Associations in...

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Complex cognition relies on both local processing and long-range brain interactions.
• Studying these interactions is crucial for understanding neural computations and representations.

Purpose of the Study:

• To investigate distinct learning signals in the lateral orbitofrontal cortex (lOFC) and ventral midbrain (VM).
• To understand how these signals interact with hippocampal representations during learning.

Main Methods:

• Coupled computational fMRI with trial-by-trial fMRI suppression.
• Measured local computations and their interaction with neural representations during learning.

Main Results:

• Identified distinct learning signals in lOFC (outcome type) and VM (reward value).
• These signals predicted trial-by-trial changes in hippocampal suppression, reflecting stimulus-outcome associations.
• Medial temporal lobe network suppression changed proportionally to stimulus-outcome associations.

Conclusions:

• lOFC and VM provide distinct learning signals that update hippocampal internal models.
• This suggests a convergence of signals for flexible representation of stimulus-outcome associations.
• Highlights the interaction between local computations and distant brain regions in learning.