Format

Send to

Choose Destination
Nat Methods. 2018 Oct;15(10):805-815. doi: 10.1038/s41592-018-0109-9. Epub 2018 Sep 17.

Inferring single-trial neural population dynamics using sequential auto-encoders.

Pandarinath C1,2,3,4,5, O'Shea DJ6,7, Collins J8,9, Jozefowicz R8,10, Stavisky SD11,6,12,7, Kao JC6,13, Trautmann EM7, Kaufman MT7,14, Ryu SI6,15, Hochberg LR16,17,18, Henderson JM11,12, Shenoy KV6,12,19,20,21,22, Abbott LF23,24,25, Sussillo D26,27,28.

Author information

1
Wallace H. Coulter Department of Biomedical Engineering, Emory University and Georgia Institute of Technology, Atlanta, GA, USA. chethan@gatech.edu.
2
Department of Neurosurgery, Emory University, Atlanta, GA, USA. chethan@gatech.edu.
3
Department of Neurosurgery, Stanford University, Stanford, CA, USA. chethan@gatech.edu.
4
Department of Electrical Engineering, Stanford University, Stanford, CA, USA. chethan@gatech.edu.
5
Stanford Neurosciences Institute, Stanford University, Stanford, CA, USA. chethan@gatech.edu.
6
Department of Electrical Engineering, Stanford University, Stanford, CA, USA.
7
Neurosciences Graduate Program, Stanford University, Stanford, CA, USA.
8
Google AI, Google Inc., Mountain View, CA, USA.
9
University of California, Berkeley, Berkeley, CA, USA.
10
OpenAI, San Francisco, CA, USA.
11
Department of Neurosurgery, Stanford University, Stanford, CA, USA.
12
Stanford Neurosciences Institute, Stanford University, Stanford, CA, USA.
13
Department of Electrical Engineering, University of California, Los Angeles, Los Angeles, CA, USA.
14
Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, USA.
15
Department of Neurosurgery, Palo Alto Medical Foundation, Palo Alto, CA, USA.
16
VA RR&D Center for Neurorestoration and Neurotechnology, Veterans Affairs Medical Center, Providence, RI, USA.
17
Center for Neurotechnology and Neurorecovery, Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA.
18
School of Engineering and Carney Institute for Brain Science, Brown University, Providence, RI, USA.
19
Department of Neurobiology, Stanford University, Stanford, CA, USA.
20
Department of Bioengineering, Stanford University, Stanford, CA, USA.
21
Bio-X Program, Stanford University, Stanford, CA, USA.
22
Howard Hughes Medical Institute, Stanford University, Stanford, CA, USA.
23
Zuckerman Mind Brain Behavior Institute, Columbia University, New York, NY, USA.
24
Department of Neuroscience, Columbia University, New York, NY, USA.
25
Department of Physiology and Cellular Biophysics, Columbia University, New York, NY, USA.
26
Department of Electrical Engineering, Stanford University, Stanford, CA, USA. sussillo@google.com.
27
Stanford Neurosciences Institute, Stanford University, Stanford, CA, USA. sussillo@google.com.
28
Google AI, Google Inc., Mountain View, CA, USA. sussillo@google.com.

Abstract

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, which is challenging due to incomplete sampling of the neural population, trial-to-trial variability, and fluctuations in action potential timing. We introduce latent factor analysis via dynamical systems, 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, latent factor analysis via dynamical systems 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.

PMID:
30224673
DOI:
10.1038/s41592-018-0109-9

Supplemental Content

Full text links

Icon for Nature Publishing Group
Loading ...
Support Center