Deep Representation Learning on Whole-Brain Population Dynamics Uncovers Geometrically Separable Neural Codes
Amina Abdelbaki, Paul Bandow, Karen Y. Cheng, Ilona C Grunwald Kadow, Martin Paul Nawrot, Vahid Rostami
The paper develops a deep-learning framework to learn interpretable representations of whole-brain neural dynamics in Drosophila without neuronal identification, revealing factorized organization along metabolic state, sensory modality, and stimulus valence axes.
bioRxiv · Sep 28, 2026 · PreprintNeuroscience