Contributed Talk Sessions | Poster Sessions | All Posters | Search Papers

Poster Session B: Wednesday, August 13, 1:00 – 4:00 pm, de Brug & E‑Hall

Learning neuronal manifolds for interacting neuronal populations

Akshey Kumar¹, Moritz Grosse-Wentrup²; ¹Universität Vienna, ²University of Vienna

Presenter: Akshey Kumar

Understanding how neuronal populations interact to process information and generate behavior is a central goal of neuroscience. However, high dimensionality, dense interactions, and unobserved factors complicate this task. The neuronal manifold hypothesis suggests that relevant dynamics occur on a lower-dimensional manifold, but it offers limited insight into the interactions among subsystems. We introduce a BunDLe-Net-based architecture that embeds distinct neuronal populations into separate latent dimensions. By leveraging BunDLe-Net’s Markovian embedding, we ensure that every point in the latent space retains predictive information about future behavioral dynamics. We apply our method to C. elegans neuronal data categorized into sensory, motor, and interneurons. The manifold not only reveals recurring motifs in the dynamics but also shows how different populations orchestrate these motifs. From the manifold, we can read off which populations encode information and drive the dynamics in each behavioral state. Thus, we present a powerful visual tool that reveals how information is processed and relayed across populations.

Topic Area: Brain Networks & Neural Dynamics

Extended Abstract: Full Text PDF