Sensitivity Derivatives for Flexible Sensorimotor Learning

TLDR

Effective learning requires knowledge of sensitivity derivatives, yet how biological systems acquire these derivatives remains unexplained and some argue they are innate. The study demonstrates that sensitivity derivatives can be learned through brain‑available information transport mechanisms. The authors propose implicit supervision, a form of information transport, to learn sensitivity derivatives and account for rapid, flexible sensorimotor adaptation in high‑dimensional contexts. Results show that sensitivity derivatives are not purely innate and that implicit supervision explains the observed flexibility and speed of sensorimotor learning.

Abstract

To learn effectively, an adaptive controller needs to know its sensitivity derivatives—the variables that quantify how system performance depends on the commands from the controller. In the case of biological sensorimotor control, no one has explained how those derivatives themselves might be learned, and some authors suggest they are not learned at all but are known innately. Here we show that this knowledge cannot be solely innate, given the adaptive flexibility of neural systems. And we show how it could be learned using forms of information transport that are available in the brain. The mechanism, which we call implicit supervision, helps explain the flexibility and speed of sensorimotor learning and our ability to cope with high-dimensional work spaces and tools.