Feedback-related negativity codes prediction error but not behavioral adjustment during probabilistic reversal...

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Electrophysiological activity in the medial frontal cortex (MFC) is crucial for behavioral adjustment.
• Previous research linked the feedback-related negativity (FRN) to prediction errors and decision value adjustments.
• Existing paradigms often conflate reward prediction error with explicit rule-based behavior.

Purpose of the Study:

• To differentiate the neural mechanisms underlying reward prediction error signaling from explicit rule-based behavioral adjustment.
• To investigate the roles of FRN and P3 amplitudes in outcome-based learning and behavioral change.

Main Methods:

• Utilized a probabilistic reversal learning task with frequent contingency reversals.
• Recorded electrophysiological activity (EEG) from the MFC.
• Employed reinforcement learning (RL) modeling to derive prediction errors.

Main Results:

• Greater FRN amplitudes correlated with larger RL model-derived prediction errors.
• Expected negative outcomes preceding rule-based behavioral reversals did not elicit an FRN.
• P3 amplitudes were significantly larger for expected negative outcomes that preceded rule-based reversals compared to unexpected negative outcomes.

Conclusions:

• The FRN specifically reflects prediction errors and associated RL-based value adjustments.
• The P3 amplitude is associated with the adjustment of behavior based on explicit rules.
• This study successfully disentangled the neural correlates of prediction error signaling and rule-based behavioral control.