Graph neural networks uncover structure and functions underlying the activity of simulated neural assemblies

TL;DR

This paper demonstrates that graph neural networks can effectively predict and interpret the activity of large simulated neural assemblies, revealing underlying connectivity, neuron types, and external stimuli.

Contribution

It introduces a novel application of graph neural networks for interpretable analysis of neural assembly dynamics, surpassing traditional methods in interpretability.

Findings

Successfully reveals connectivity matrices and neuron types

Accurately predicts neural activity dynamics

Provides interpretable insights into neural mechanisms

Abstract

Graph neural networks trained to predict observable dynamics can be used to decompose the temporal activity of complex heterogeneous systems into simple, interpretable representations. Here we apply this framework to simulated neural assemblies with thousands of neurons and demonstrate that it can jointly reveal the connectivity matrix, the neuron types, the signaling functions, and in some cases hidden external stimuli. In contrast to existing machine learning approaches such as recurrent neural networks and transformers, which emphasize predictive accuracy but offer limited interpretability, our method provides both reliable forecasts of neural activity and interpretable decomposition of the mechanisms governing large neural assemblies.