Joint Learning of Discrete and Continuous Variability with Coupled Autoencoding Agents

Jointly identifying discrete and continuous factors of variability can help unravel complex phenomena. In neuroscience, a high-priority instance of this problem is the analysis of neuronal identity. Here, we study this problem in a variational framework by utilizing interacting autoencoding agents, designed to function in the absence of prior distribution over the discrete variable and reach collective decisions. We provide theoretical justification for our method and demonstrate improvements in terms of interpretability, stability, and accuracy over comparable approaches with experimental results on two benchmark datasets and a recent dataset of gene expression profiles of mouse cortical neurons. Furthermore, we demonstrate how our method can determine the neuronal cell types in an unsupervised setting, while identifying the genes implicated in regulating biologically relevant neuronal states.

A coupled autoencoder approach for multi-modal analysis of cell types

Recent developments in high throughput profiling of individual neurons have spurred data driven exploration of the idea that there exist natural groupings of neurons referred to as cell types. The promise of this idea is that the immense complexity of brain circuits can be reduced, and effectively studied by means of interactions between cell types. While clustering of neuron populations based on a particular data modality can be used to define cell types, such definitions are often inconsistent across different characterization modalities. We pose this issue of cross-modal alignment as an optimization problem and develop an approach based on coupled training of autoencoders as a framework for such analyses. We apply this framework to a Patch-seq dataset consisting of transcriptomic and electrophysiological profiles for the same set of neurons to study consistency of representations across modalities, and evaluate cross-modal data prediction ability. We explore the problem where only a subset of neurons is characterized with more than one modality, and demonstrate that representations learned by coupled autoencoders can be used to identify types sampled only by a single modality.