Goal-Directed Behavior and Instrumental Devaluation: A Neural System-Level Computational Model.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Behavioral Neuroscience

Background:

• Devaluation experiments are crucial for understanding goal-directed behaviors in mammals.
• The neural basis for how current reward value influences instrumental actions remains incompletely understood.

Purpose of the Study:

• To propose a neural system-level computational model explaining how reward value controls instrumental actions.
• To hypothesize the integrated brain mechanisms underlying devaluation effects on goal-directed behavior.

Main Methods:

• Developed a computational model based on the hypothesis that internal representations of actions activate reward representations with state-dependent values.
• Integrated key brain components: basolateral amygdala-insular cortex, basal ganglia-cortical loops, and cortico-cortical/striato-nigro-striatal pathways.
• Validated the model by reproducing and explaining results from rat devaluation experiments with specific brain lesions.

Main Results:

• The model demonstrates computational soundness in linking action-outcome associations and state-dependent reward valuation.
• It successfully explains how various brain structures (amygdala, insular cortex, basal ganglia, cortex) interact to bias action selection.
• The model accurately reproduces lesion and control data from devaluation experiments.

Conclusions:

• The proposed model provides a systems-level framework for understanding the neural mechanisms of devaluation.
• It highlights the integrated roles of specific brain circuits in attributing current value to rewards and guiding actions.
• The model serves as a foundation for future research on motivational control of goal-directed behavior.