Extended Data Fig. 8: Representation and decoding of neural signals during macaque arm-reaching kinematics using supervised targeted dimensionality reduction (TDR).
From: MARBLE: interpretable representations of neural population dynamics using geometric deep learning
This method regresses neural firing rates to subspaces defined by the different reaching directions. a Ground truth hand trajectories of a macaque in seven reach conditions. b Latent representation of neural data across conditions in a single session regressed to the corresponding TDR subspaces. c Linear decoding of hand trajectories from latent representations.
