Predictive decision making driven by multiple time-linked reward representations in the anterior cingulate cortex

Area of Science:

• Neuroscience
• Cognitive Science
• Decision Making

Background:

• Environmental factors dynamically alter choice values.
• Predicting future choice values relies on discerning these trends.
• Understanding the neural basis of trend tracking is crucial for decision-making research.

Purpose of the Study:

• To investigate how humans track changing trends in reward rates.
• To identify the neural mechanisms underlying trend-based decision-making.
• To model the computational processes involved in trend-guided choices.

Main Methods:

• Functional neuroimaging to observe brain activity during decision tasks.
• Behavioral analysis of choice persistence and switching.
• Computational modeling using reinforcement-learning principles.

Main Results:

• The dorsal anterior cingulate cortex (dACC) simultaneously holds estimates of recent and past reward rates.
• Opposing signals in dACC reflect comparisons of reward rates with decision weights.
• The strength of these representations predicts behavioral persistence or switching.

Conclusions:

• The dACC plays a key role in tracking environmental trends for adaptive decision-making.
• Reinforcement learning models can capture trend-guided choice behavior.
• Distinct brain regions (dACC, ventral striatum, ventromedial prefrontal cortex) are involved in processing expected and instantaneous prediction errors and choice signals.