Disentangling the flow of signals between populations of neurons

Abstract

Technological advances have granted the ability to record from large populations of neurons across multiple brain areas. These recordings may illuminate how communication between areas contributes to brain function, yet a substantial barrier remains: How do we disentangle the concurrent, bidirectional flow of signals between two populations of neurons? We therefore propose here a novel dimensionality reduction framework: Delayed Latents Across Groups (DLAG). DLAG disentangles signals relayed in both directions; identifies how these signals are represented by each population; and characterizes how they evolve over time and trial-to-trial. We demonstrate that DLAG performs well on synthetic datasets similar in scale to current neurophysiological recordings. Then we study simultaneously recorded populations in primate visual areas V1 and V2, where DLAG reveals signatures of concurrent yet selective communication. Our framework lays the foundation for dissecting the intricate flow of signals across populations of neurons, and how this signaling contributes to cortical computation.

An Efficient Spectral Trust-Region Deflation Method for Multiple Solutions

“…Hultman et al 2018 examined multiregion local field potential data and identified frequencybased interactions across brain regions using a Gaussian Process Factor Analysis model. Following a similar approach, Gokcen et al 2022;2023 suggested that latent variables can be divided into across-and within-region components. This model was applied to disentangle the concurrent and bidirectional communications across brain regions with a multi-output Squared Exponential kernel.…”