On the choice of metric in gradient-based theories of brain function

TL;DR

This paper reviews the mathematical basis of gradient descent in brain function models, highlighting how the choice of metric affects predictions and proposing ways to better constrain this choice for more accurate models.

Contribution

It clarifies the importance of metric choice in gradient-based brain models and suggests methods to constrain this choice for improved predictive power.

Findings

Gradient descent's effectiveness depends on metric choice.

Common pitfalls arise from arbitrary metric selection.

Proposes methods to constrain the metric in models.

Abstract

The idea that the brain functions so as to minimize certain costs pervades theoretical neuroscience. Since a cost function by itself does not predict how the brain finds its minima, additional assumptions about the optimization method need to be made to predict the dynamics of physiological quantities. In this context, steepest descent (also called gradient descent) is often suggested as an algorithmic principle of optimization potentially implemented by the brain. In practice, researchers often consider the vector of partial derivatives as the gradient. However, the definition of the gradient and the notion of a steepest direction depend on the choice of a metric. Since the choice of the metric involves a large number of degrees of freedom, the predictive power of models that are based on gradient descent must be called into question, unless there are strong constraints on the choice of the metric. Here we provide a didactic review of the mathematics of gradient descent, illustrate common pitfalls of using gradient descent as a principle of brain function with examples from the literature and propose ways forward to constrain the metric.