Abstract
Despite rapid advances in machine learning tools, the majority of neural decoding approaches still use traditional methods. Modern machine learning tools, which are versatile and easy to use, have the potential to significantly improve decoding performance. This tutorial describes how to effectively apply these algorithms for typical decoding problems. We provide descriptions, best practices, and code for applying common machine learning methods, including neural networks and gradient boosting. We also provide detailed comparisons of the performance of various methods at the task of decoding spiking activity in motor cortex, somatosensory cortex, and hippocampus. Modern methods, particularly neural networks and ensembles, significantly outperform traditional approaches, such as Wiener and Kalman filters. Improving the performance of neural decoding algorithms allows neuroscientists to better understand the information contained in a neural population and can help to advance engineering applications such as brain–machine interfaces. Our code package is available at github.com/kordinglab/neural_decoding.
- Neural decoding
- Machine learning
- Neural data analysis
- Deep learning
- Motor cortex
- Somatosensory cortex
- Hippocampus
Footnotes
The authors declare no competing financial interests.
J.I.G. was supported by National Institutes of Health (NIH) Grants F31-EY-025532 and T32-HD-057845, National Science Foundation NeuroNex Award DBI-1707398, and the Gatsby Charitable Foundation. A.S.B. was supported by NIH Grant MH-103910. M.G.P. was supported by NIH Grants F31-NS-092356 and T32-HD-07418. R.H.C. was supported by NIH Grants R01-NS-095251 and DGE-1324585. L.E.M. was supported by NIH Grants R01-NS-074044 and R01-NS-095251. K.P.K. was supported by NIH Grants R01-NS-074044, R01-NS-063399, and R01-EY-021579.
This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license, which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.