Theory-experiment cycle for understanding intelligence: An example of the decision neuroscience and reinforcement learning

Abstract

Cognitive science is a framework for understanding human behavior using the metaphor of a computational machine. Computational neuroscience has also taken the approach of using mathematical algorithms to reveal the computational mechanisms of the brain. In this paper, we review an approach to reveal the computational mechanisms of the brain using reinforcement learning to explain behaviors, especially those related to reward learning and decision making, and its implications for the surrounding fields. Computational modeling with reinforcement learning provides a novel way of understanding and applications not only in neuroscience but also in various surrounding fields such as psychology, economics, marketing, and psychiatry. Finally, we will discuss the limitations of the mathematical approach to understanding the brain and the future direction of cognitive science.