Flow matching unifies measures of brain response differences
A Flow Matching Framework for Neural Representational Dissimilarity
Artificial IntelligenceInformation TheoryMachine Learning
Summary
Measuring how brain activity differs between conditions or models helps scientists understand how the brain works. The authors show that various ways to measure these differences can be seen as part of one general method called flow matching. This approach can estimate differences more accurately, especially when brain responses are complicated or vary continuously. It also allows creating new measures in a systematic way.
What this means in practice
- •For neuroimaging data analysts: Estimate differences between neural response patterns more accurately for complex brain data using flow matching based metrics.
- •For machine learning engineers: Design new similarity measures for comparing outputs of neural networks based on flow matching principles.
Authors
Zeyuan Ye, Xue-Xin Wei
Abstract
Neural representational dissimilarity quantifies differences between neural response distributions, and is essential for comparing neural codes across stimuli, brain areas, tasks, and models. Commonly used distance metrics involve different assumptions and are estimated with separate methods. Here, we show that a variety of distance metrics can be unified under a flow matching framework developed in deep generative models. That is, these distances arise as Jeffreys divergences under different velocity constraints. We find that flow matching has advantages for estimating distances involving complicated distributions and continuous variables. Furthermore, this framework enables the design of new distance metrics in a principled way. Together, flow matching provides a unified approach for understanding, estimating, and designing neural representational dissimilarity metrics.