Distinguishing Learning Rules with Brain Machine Interfaces

TL;DR

This paper proposes a method to differentiate biologically plausible supervised and reinforcement learning rules in neural networks by analyzing activity changes during learning, using brain-machine interface experiments for validation.

Contribution

It introduces a metric to distinguish learning rules based on activity changes, leveraging BMI experiments to identify the underlying learning mechanism in neural models.

Findings

The metric successfully differentiates learning rules in simulated BMI tasks.

Reinforcement learning aligns more closely with true gradient updates.

Supervised learning shows bias due to imperfect credit assignment modeling.

Abstract

Despite extensive theoretical work on biologically plausible learning rules, clear evidence about whether and how such rules are implemented in the brain has been difficult to obtain. We consider biologically plausible supervised- and reinforcement-learning rules and ask whether changes in network activity during learning can be used to determine which learning rule is being used. Supervised learning requires a credit-assignment model estimating the mapping from neural activity to behavior, and, in a biological organism, this model will inevitably be an imperfect approximation of the ideal mapping, leading to a bias in the direction of the weight updates relative to the true gradient. Reinforcement learning, on the other hand, requires no credit-assignment model and tends to make weight updates following the true gradient direction. We derive a metric to distinguish between learning rules by observing changes in the network activity during learning, given that the mapping from brain to behavior is known by the experimenter. Because brain-machine interface (BMI) experiments allow for precise knowledge of this mapping, we model a cursor-control BMI task using recurrent neural networks, showing that learning rules can be distinguished in simulated experiments using only observations that a neuroscience experimenter would plausibly have access to.