The promise and perils of neural manifolds

A talk by Devika Narain (Associate Professor of Radboud University in Nijmegen, Netherlands)

Neural manifolds have emerged as a powerful framework for understanding neural population activity, revealing geometric and topological structure hidden within high-dimensional neural state spaces. In this workshop, we will trace the conceptual development of neural manifolds from population dynamics and dimensionality reduction to modern geometric approaches. We will examine what neural manifolds are, how they can be recovered from neural data, and when linear methods provide adequate descriptions of neural activity versus when nonlinear manifold learning approaches become necessary. Particular emphasis will be placed on the role of geometry, topology, and distortion in the faithful reconstruction and interpretation of neural manifolds in neurophysiological data.

If you want to speak with Devika, please contact bkuhn@oist.jp