Optimizing the depth and the direction of prospective planning using information values.

Area of Science:

• Cognitive Science
• Computational Neuroscience
• Behavioral Economics

Background:

• Evaluating future consequences is crucial for decision-making.
• Deep search tree expansion is computationally expensive.
• Plan-until-habit schemes use limited depth and habitual values.

Purpose of the Study:

• To address key questions in search tree expansion: direction and stopping criteria.
• To propose a principled algorithm balancing planning speed and accuracy.
• To explain and predict behavioral patterns in decision-making.

Main Methods:

• Developed a novel algorithm based on a speed/accuracy tradeoff for search tree expansion.
• Simulated the algorithm's performance in a grid-world environment.
• Validated the algorithm against known animal and human behavioral patterns.

Main Results:

• The algorithm effectively and efficiently expands search trees.
• It explains the impact of time pressure and reward magnitude on planning.
• Demonstrates the shift from goal-directed to habitual behavior with training.

Conclusions:

• The proposed algorithm offers a principled solution for efficient future planning.
• It provides a unified framework for understanding diverse decision-making behaviors.
• The algorithm generates testable predictions for future experimental validation.