AUTHOR=Bossaerts Peter 

TITLE=Formalizing the Function of Anterior Insula in Rapid Adaptation

JOURNAL=Frontiers in Integrative Neuroscience

VOLUME=Volume 12 - 2018

YEAR=2018

URL=https://www.frontiersin.org/journals/integrative-neuroscience/articles/10.3389/fnint.2018.00061

DOI=10.3389/fnint.2018.00061

ISSN=1662-5145

ABSTRACT=Anterior Insula (AI) is thought to play a crucial role in rapid adaptation in an ever-changing environment. Mathematically, it is known to track risk and surprise (unsigned prediction errors). Modern theories of learning, however, assign a dominant role to signed prediction errors, not to risk and surprise. Risk and surprise only enter to the extent that they modulate the learning rate, and even without such modulation, adaptation will still be effective, albeit slow. Here, I propose a new theory of learning, Reference-Model Based Learning (RMBL), where risk and surprise are central, and prediction errors play a secondary, though still crucial, role. Choices in a target location prediction task where participants were continuously required to adapt appeared to be consistent with model predictions. AI reaction to surprise was more acute in the more difficult treatment, consistent with its hypothesized role in metacognition. I discuss links with related theories, such as Active Inference, Actor-Critic Models and Reference-Model Based Adaptive Control.