A Disentangled Low-Rank RNN Framework for Uncovering Neural Connectivity and Dynamics

TL;DR

This paper introduces DisRNN, a novel low-rank RNN framework that enhances the interpretability and disentanglement of neural latent dynamics by enforcing group-wise independence, improving understanding of neural connectivity.

Contribution

DisRNN extends low-rank RNNs by incorporating a disentanglement mechanism through a variational autoencoder framework and partial correlation penalties, enabling clearer interpretation of neural dynamics.

Findings

DisRNN outperforms baseline lrRNNs in disentanglement and interpretability.

DisRNN effectively captures complex neural dynamics in synthetic and real data.

Enhanced interpretability aids understanding of neural connectivity and computation.

Abstract

Low-rank recurrent neural networks (lrRNNs) are a class of models that uncover low-dimensional latent dynamics underlying neural population activity. Although their functional connectivity is low-rank, it lacks disentanglement interpretations, making it difficult to assign distinct computational roles to different latent dimensions. To address this, we propose the Disentangled Recurrent Neural Network (DisRNN), a generative lrRNN framework that assumes group-wise independence among latent dynamics while allowing flexible within-group entanglement. These independent latent groups allow latent dynamics to evolve separately, but are internally rich for complex computation. We reformulate the lrRNN under a variational autoencoder (VAE) framework, enabling us to introduce a partial correlation penalty that encourages disentanglement between groups of latent dimensions. Experiments on synthetic, monkey M1, and mouse voltage imaging data show that DisRNN consistently improves the disentanglement and interpretability of learned neural latent trajectories in low-dimensional space and low-rank connectivity over baseline lrRNNs that do not encourage partial disentanglement.