Dynamic Integration of Value Information into a Common Probability Currency as a Theory for Flexible Decision Making

Jill X O’Reilly,Vassilios Christopoulos,Paul R Schrater

doi:10.1371/journal.pcbi.1004402

Jill X O’Reilly, Vassilios Christopoulos + Show 1 more

Open Access

https://doi.org/10.1371/journal.pcbi.1004402

Copy DOI

Abstract

Decisions involve two fundamental problems, selecting goals and generating actions to pursue those goals. While simple decisions involve choosing a goal and pursuing it, humans evolved to survive in hostile dynamic environments where goal availability and value can change with time and previous actions, entangling goal decisions with action selection. Recent studies suggest the brain generates concurrent action-plans for competing goals, using online information to bias the competition until a single goal is pursued. This creates a challenging problem of integrating information across diverse types, including both the dynamic value of the goal and the costs of action. We model the computations underlying dynamic decision-making with disparate value types, using the probability of getting the highest pay-off with the least effort as a common currency that supports goal competition. This framework predicts many aspects of decision behavior that have eluded a common explanation.

Highlights

A soccer player moves the ball down the field, looking for an open teammate or a chance to score a goal
This information is diverse relating to both the dynamic value of the goal and the dynamic action cost, creating a challenging problem in integrating information across these diverse types in real time
When there are more targets in one hemifield than the other, there are more alternative reaching policies towards this space biasing the competition to that side, Fig 4H. These findings show that weighting individual policies with the relative desirability values can explain many aspects of human behavior in reaching decisions with competing goals

Summary

Introduction

A soccer player moves the ball down the field, looking for an open teammate or a chance to score a goal. He/she must select between many competing goals while acting, whose costs and benefits can change dynamically during ongoing actions. In this game scenario, the attacker has options to pass the ball to one of his/her teammates. To decide which strategy to follow at a given moment requires dynamically integrating value information from disparate sources. This information is diverse relating to both the dynamic value of the goal (i.e., relative reward of the goal, probability that reward is available for that goal) and the dynamic action cost (i.e., cost of actions to pursue that goal, precision required), creating a challenging problem in integrating information across these diverse types in real time. Despite intense research in decision neuroscience, dynamic value integration into a common currency remains poorly understood

Methods

Results

Discussion

Conclusion