The neural coding of expected and unexpected monetary performance outcomes: dissociations between active and...

Area of Science:

• Neuroscience
• Cognitive Science
• Decision Making

Background:

• Environmental adaptation relies on learning stimulus-response-outcome associations.
• Learning can occur actively through trial and error or passively through observation.
• Understanding the neural basis of these distinct learning modes is crucial.

Purpose of the Study:

• To investigate similarities and differences in neural mechanisms between active and observational learning from monetary feedback.
• To compare brain activation patterns associated with expected and unexpected monetary outcomes in both learning types.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to study two groups of 15 participants each: active learners and observational learners.
• Active learners made choices and received monetary feedback; observational learners watched active learners' choices and outcomes.
• Learning performance was assessed using feedback-free test trials.

Main Results:

• Comparable learning performance was observed between active and observational learners.
• Distinct neural activation patterns emerged for processing unexpected versus expected monetary feedback, particularly for positive outcomes.
• Active learning showed greater right striatum activity for unexpected rewards, while observational learning exhibited enhanced bilateral hippocampal activity.
• Prediction error (PE) modulation was seen in the right putamen for both groups, with stronger PE-related activations in the right anterior caudate nucleus and medial orbitofrontal cortex for active learning.

Conclusions:

• The striatum and orbitofrontal cortex are more involved in linking reward stimuli to one's own actions compared to observing another person's actions.
• Neural mechanisms underlying active and observational learning from feedback show both shared and distinct patterns, particularly in reward processing and prediction error computation.