Predictive information in a sensory population

TLDR

Prediction is essential for life, yet it is unclear whether humans and brains are close to optimal predictors. The study aims to assess prediction efficiency by measuring the information neurons carry about future sensory experiences. The authors quantify the predictive information in retinal neurons to evaluate this efficiency. They find that retinal neuron groups efficiently separate predictive from nonpredictive information, supporting efficient coding as a general principle across neural computation.

Abstract

Significance Prediction is an essential part of life. However, are we really “good” at making predictions? More specifically, are pieces of our brain close to being optimal predictors? To assess the efficiency of prediction, we need to measure the information that neurons carry about the future of our sensory experiences. We show how to do this, at least in simplified contexts, and find that groups of neurons in the retina indeed are close to maximally efficient at separating predictive information from the nonpredictive background. Efficient coding of predictive information is a principle that can be applied at every stage of neural computation.