AI Insight
Researchers developed MR-LFADS, a machine learning model that analyzes recordings from multiple brain regions simultaneously to distinguish between three key factors: direct communication between regions, inputs from unobserved brain areas, and local neural activity within each region. The model outperformed existing methods in simulated neural networks and successfully predicted the effects of experimental brain circuit manipulations in real electrophysiology data that weren't used during model training. This approach addresses a fundamental challenge in neuroscience: accurately identifying how different brain regions influence each other from complex multi-region recordings.
Why it matters
This tool could accelerate discovery of brain-wide communication principles and help researchers understand how distributed neural circuits coordinate during behavior and cognition. The ability to predict effects of circuit perturbations suggests potential applications in designing targeted interventions for neurological disorders.
arXiv:2506.19094v5 Announce Type: replace
Abstract: Neural recording technologies now enable simultaneous recording of population activity across many brain regions, motivating the development of data-driven models of inter-regional communication. However, existing models can struggle to disentangle the influences that drive recorded population activity, leading to inaccurate portraits of communication. Here, we introduce Multi-Region Latent Factor Analysis via Dynamical Systems (MR-LFADS), a sequential variational autoencoder designed to disentangle inter-regional communication, inputs from unobserved regions, and local neural population dynamics. We show that MR-LFADS outperforms existing approaches at identifying communication across dozens of simulations of task-trained multi-region networks. When applied to large-scale electrophysiology, MR-LFADS predicts brain-wide effects of circuit perturbations that were held out during model fitting. These validations on synthetic and real neural data position MR-LFADS as a promising tool for discovering principles of brain-wide information processing.
Source: Accurate identification of communication between multiple interacting neural populations