Vicarious reinforcement learning signals when instructing others.

Area of Science:

• Neuroscience
• Cognitive Science
• Social Learning

Background:

• Reinforcement learning (RL) theory explains learning via prediction errors (PEs).
• Social learning often involves teachers providing feedback that generates PEs in students.
• The neural basis of RL signals in teachers monitoring others is largely unknown.

Purpose of the Study:

• To investigate the neural mechanisms underlying reinforcement learning signals in teachers.
• To determine if the anterior cingulate cortex (ACC) signals prediction errors when teachers monitor student learning.
• To explore other brain regions involved in processing student performance during teaching.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to scan teachers.
• Teachers instructed confederate students on action-outcome associations, providing feedback.
• A reinforcement learning computational model analyzed student behavior and generated prediction error values.

Main Results:

• Teacher ACC activity significantly correlated with student prediction error values.
• Insula and ventromedial prefrontal cortex activity in teachers reflected student predicted values.
• These findings indicate vicarious processing of prediction errors in teachers.

Conclusions:

• The anterior cingulate cortex (ACC) plays a role in signaling vicarious prediction errors during teaching.
• Reinforcement learning mechanisms, when processed vicariously, may underlie and facilitate teaching behaviors.
• This study provides novel insights into the neural basis of social learning and instruction.