Event:

It’s not only the lack of information but also the abundance of information that poses challenges to the brain: bombarded with a flurry of bits of information the brain needs to identify the few relevant pieces that are useful for efficient prediction of future experiences and for devising optimal actions. In this talk I will discuss several aspects of the computations necessary for discovering relevance of environmental features. First, I will discuss memory experiment to address how the distortion function in rate distortion theory reveals the features relevant for encoding information. Next, I will discuss how optimality principles help humans navigate the space of features, eventually leading to deeming the experimental instructions irrelevant. Finally I will discuss how experimentally relevant bits of information can be efficiently represented during task execution in mice, and computations critical for managing the changing relevance of features can be identified in the Anterior Cingulate Cortex of mice.