Prospective and retrospective values integrated in frontal cortex drive predictive choice.

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• The brain must constantly update action-values in dynamic environments.
• Action-values are derived from past experiences (retrospective) and environmental knowledge (prospective).
• The neural circuits integrating these values for action execution remain largely unknown.

Purpose of the Study:

• To investigate how the brain integrates retrospective and prospective action-values.
• To identify the neural circuits responsible for knowledge-dependent predictive choice behavior.

Main Methods:

• Utilized reinforcement learning and two-photon calcium imaging in mice.
• Combined computational modeling with in vivo neural recordings.
• Employed optogenetic manipulation to probe neural circuit function.

Main Results:

• Preparatory neural activity in the anterior-lateral motor (ALM) area initially encoded retrospective value.
• After extensive training, ALM neurons jointly encoded both retrospective and prospective values.
• Inhibiting ALM preparatory activity disrupted expert predictive choice behavior, reverting mice to a novice-like state.

Conclusions:

• The integrated action-value representation in ALM preparatory activity is crucial for predictive choice behavior.
• ALM plays a key role in biasing actions based on learned environmental knowledge.
• This study elucidates a critical neural mechanism for adaptive decision-making in volatile environments.