Latent Graph Learning in Generative Models of Neural Signals

TL;DR

This paper investigates how to extract interpretable latent graph structures from generative models of neural signals, revealing partial alignment with true neural connectivity and suggesting new directions for model constraints.

Contribution

It introduces methods for latent graph learning in neural generative models and evaluates their alignment with ground-truth neural connectivity.

Findings

Modest alignment between learned and true directed graphs.

Strong alignment in co-input graph representations.

Results motivate incorporating graph constraints in models.

Abstract

Inferring temporal interaction graphs and higher-order structure from neural signals is a key problem in building generative models for systems neuroscience. Foundation models for large-scale neural data represent shared latent structures of neural signals. However, extracting interpretable latent graph representations in foundation models remains challenging and unsolved. Here we explore latent graph learning in generative models of neural signals. By testing against numerical simulations of neural circuits with known ground-truth connectivity, we evaluate several hypotheses for explaining learned model weights. We discover modest alignment between extracted network representations and the underlying directed graphs and strong alignment in the co-input graph representations. These findings motivate paths towards incorporating graph-based geometric constraints in the construction of large-scale foundation models for neural data.