Go and no-go learning in reward and punishment: interactions between affect and effect

Area of Science:

• Neuroscience
• Cognitive Science
• Decision Science

Background:

• Decision-making is governed by affect (valence) and effect (action).
• Understanding how humans learn instrumental actions under different outcome valences is crucial.
• Existing models often struggle to disentangle instrumental and Pavlovian influences.

Purpose of the Study:

• To investigate the acquisition of instrumental responding when action requirements and outcome valence are orthogonalized.
• To differentiate instrumental and Pavlovian contributions to learning asymmetries.
• To identify neural correlates of instrumental learning using model-based fMRI.

Main Methods:

• Behavioral experiments with healthy volunteers using a task with orthogonalized action and valence.
• Computational reinforcement-learning models to analyze learning components.
• Model-based functional Magnetic Resonance Imaging (fMRI) to examine brain activity.

Main Results:

• Subjects learned active choices better in rewarded conditions and passive choices in punished conditions.
• Computational models revealed instrumental learning is modulated by Pavlovian biases.
• BOLD signals in the striatum and substantia nigra/ventral tegmental area (SN/VTA) correlated with action values.
• Successful instrumental learning engaged the bilateral inferior frontal gyrus.

Conclusions:

• Instrumental learning is contingent on overcoming inherent and plastic Pavlovian biases.
• Neural activity in the striatum, SN/VTA, and inferior frontal gyrus underpins instrumental learning.
• This learning involves distinct brain activity patterns related to action and inhibition.