Litcius/Paper detail

MARBLE: interpretable representations of neural population dynamics using geometric deep learning

Adam Gosztolai, Robert L. Peach, Alexis Arnaudon, Mauricio Barahona, Pierre Vandergheynst

2025Nature Methods18 citationsDOIOpen Access PDF

Abstract

The dynamics of neuron populations commonly evolve on low-dimensional manifolds. Thus, we need methods that learn the dynamical processes over neural manifolds to infer interpretable and consistent latent representations. We introduce a representation learning method, MARBLE, which decomposes on-manifold dynamics into local flow fields and maps them into a common latent space using unsupervised geometric deep learning. In simulated nonlinear dynamical systems, recurrent neural networks and experimental single-neuron recordings from primates and rodents, we discover emergent low-dimensional latent representations that parametrize high-dimensional neural dynamics during gain modulation, decision-making and changes in the internal state. These representations are consistent across neural networks and animals, enabling the robust comparison of cognitive computations. Extensive benchmarking demonstrates state-of-the-art within- and across-animal decoding accuracy of MARBLE compared to current representation learning approaches, with minimal user input. Our results suggest that a manifold structure provides a powerful inductive bias to develop decoding algorithms and assimilate data across experiments.

Topics & Concepts

Artificial intelligenceManifold (fluid mechanics)Artificial neural networkRepresentation (politics)Computer scienceDecoding methodsNonlinear dimensionality reductionModels of neural computationDynamical systems theoryUnsupervised learningComputationPattern recognition (psychology)Machine learningAlgorithmDimensionality reductionPhysicsMechanical engineeringQuantum mechanicsPoliticsPolitical scienceLawEngineeringNeural dynamics and brain functionNeural Networks and ApplicationsAdvanced Memory and Neural Computing
MARBLE: interpretable representations of neural population dynamics using geometric deep learning | Litcius