How much neuroscience does a neuroscientist need to know?

TL;DR

This paper argues that understanding key biological constraints is crucial for modeling brain algorithms, which helps predict neural responses and bridges computational neuroscience with AI interpretability.

Contribution

It introduces the idea that simple biological details constrain plausible algorithms, enabling interpretation of neural responses and unifying neuroscience with AI mechanistic interpretability.

Findings

Biological details like nonnegative firing influence model algorithms.

Models with biological constraints predict single-neuron responses accurately.

Symmetry-breaking in models leads to interpretable neural responses.

Abstract

How much of the brain's learned algorithms depend on the fact it is a brain? We argue: a lot, but surprisingly few details matter. We point to simple biological details -- e.g. nonnegative firing and energetic/space budgets in connectionist architectures -- which, when mixed with the requirements of solving a task, produce models that predict brain responses down to single-neuron tuning. We understand this as details constraining the set of plausible algorithms, and their implementations, such that only `brain-like' algorithms are learned. In particular, each biological detail breaks a symmetry in connectionist models (scale, rotation, permutation) leading to interpretable single-neuron responses that are meaningfully characteristic of particular algorithms. This view helps us not only understand the brain's choice of algorithm but also infer algorithm from measured neural responses. Further, this perspective aligns computational neuroscience with mechanistic interpretability in AI, suggesting a more unified approach to studying the mechanisms of intelligence, both natural and artificial.