Coulter Department of Biomedical Engineering
Georgia Institute of Technology & Emory University School of Medicine
Neuroscience is experiencing a revolution in which simultaneous recording of thousands of neurons is revealing population dynamics that are not apparent from single-neuron responses. This structure is typically extracted from data averaged across many trials, but deeper understanding requires studying phenomena detected in single trials. Single-trial characterization is challenging due to incomplete sampling of the neural population, trial-to-trial variability, and fluctuations in action potential timing. I will discuss a method we recently developed, latent factor analysis via dynamical systems (LFADS), a deep learning method to infer latent dynamics from single-trial neural spiking data. When applied to a variety of macaque and human motor cortical datasets, LFADS accurately predicts observed behavioral variables, extracts precise firing rate estimates of neural dynamics on single trials, infers perturbations to those dynamics that correlate with behavioral choices, and combines data from non-overlapping recording sessions spanning months to improve inference of underlying dynamics. I will also discuss recent innovations that allow LFADS to be applied more broadly, to data from a variety of brain areas and behavioral paradigms, without requiring machine learning expertise. Finally, I will discuss how these approaches open new avenues to developing robust brain-machine interfaces to help people with paralysis.