A sliced inverse regression (SIR) decoding the forelimb movement from neuronal spikes in the rat motor cortex

Shih-Hung Yang, You Yin Chen, Sheng Huang Lin, Lun De Liao, Henry Horng Shing Lu, Ching Fu Wang, Po Chuan Chen, Yu Chun Lo, Thanh Dat Phan, Hsiang Ya Chao, Ching Hui Lin, Hsin Yi Lai, Wei Chen Huang

Research output: Contribution to journalArticle

Abstract

Several neural decoding algorithms have successfully converted brain signals into commands to control a computer cursor and prosthetic devices. A majority of decoding methods, such as population vector algorithms (PVA), optimal linear estimators (OLE), and neural networks (NN), are effective in predicting movement kinematics, including movement direction, speed and trajectory but usually require a large number of neurons to achieve desirable performance. This study proposed a novel decoding algorithm even with signals obtained from a smaller numbers of neurons. We adopted sliced inverse regression (SIR) to predict forelimb movement from single-unit activities recorded in the rat primary motor (M1) cortex in a water-reward lever-pressing task. SIR performed weighted principal component analysis (PCA) to achieve effective dimension reduction for nonlinear regression. To demonstrate the decoding performance, SIR was compared to PVA, OLE, and NN. Furthermore, PCA and sequential feature selection (SFS) which are popular feature selection techniques were implemented for comparison of feature selection effectiveness. Among SIR, PVA, OLE, PCA, SFS, and NN decoding methods, the trajectories predicted by SIR (with a root mean square error, RMSE, of 8.47 ± 1.32 mm) was closer to the actual trajectories compared with those predicted by PVA (30.41 ± 11.73 mm), OLE (20.17 ± 6.43 mm), PCA (19.13 ± 0.75 mm), SFS (22.75 ± 2.01 mm), and NN (16.75 ± 2.02 mm). The superiority of SIR was most obvious when the sample size of neurons was small. We concluded that SIR sorted the input data to obtain the effective transform matrices for movement prediction, making it a robust decoding method for conditions with sparse neuronal information.

Original languageEnglish
Article number556
JournalFrontiers in Neuroscience
Volume10
Issue numberDEC
DOIs
Publication statusPublished - 2016

Fingerprint

Forelimb
Motor Cortex
Principal Component Analysis
Neurons
Population
Reward
Biomechanical Phenomena
Sample Size
Equipment and Supplies
Water
Brain

Keywords

  • Forelimb movement prediction
  • Neural decoding
  • Neural networks (NN)
  • Principle component analysis (PCA)
  • Sliced inverse regression (SIR)

ASJC Scopus subject areas

  • Neuroscience(all)

Cite this

Yang, S-H., Chen, Y. Y., Lin, S. H., Liao, L. D., Lu, H. H. S., Wang, C. F., ... Huang, W. C. (2016). A sliced inverse regression (SIR) decoding the forelimb movement from neuronal spikes in the rat motor cortex. Frontiers in Neuroscience, 10(DEC), [556]. https://doi.org/10.3389/fnins.2016.00556

A sliced inverse regression (SIR) decoding the forelimb movement from neuronal spikes in the rat motor cortex. / Yang, Shih-Hung; Chen, You Yin; Lin, Sheng Huang; Liao, Lun De; Lu, Henry Horng Shing; Wang, Ching Fu; Chen, Po Chuan; Lo, Yu Chun; Phan, Thanh Dat; Chao, Hsiang Ya; Lin, Ching Hui; Lai, Hsin Yi; Huang, Wei Chen.

In: Frontiers in Neuroscience, Vol. 10, No. DEC, 556, 2016.

Research output: Contribution to journalArticle

Yang, S-H, Chen, YY, Lin, SH, Liao, LD, Lu, HHS, Wang, CF, Chen, PC, Lo, YC, Phan, TD, Chao, HY, Lin, CH, Lai, HY & Huang, WC 2016, 'A sliced inverse regression (SIR) decoding the forelimb movement from neuronal spikes in the rat motor cortex', Frontiers in Neuroscience, vol. 10, no. DEC, 556. https://doi.org/10.3389/fnins.2016.00556
Yang, Shih-Hung ; Chen, You Yin ; Lin, Sheng Huang ; Liao, Lun De ; Lu, Henry Horng Shing ; Wang, Ching Fu ; Chen, Po Chuan ; Lo, Yu Chun ; Phan, Thanh Dat ; Chao, Hsiang Ya ; Lin, Ching Hui ; Lai, Hsin Yi ; Huang, Wei Chen. / A sliced inverse regression (SIR) decoding the forelimb movement from neuronal spikes in the rat motor cortex. In: Frontiers in Neuroscience. 2016 ; Vol. 10, No. DEC.
@article{eb73c74e34f44757ad12d172a6caad4a,
title = "A sliced inverse regression (SIR) decoding the forelimb movement from neuronal spikes in the rat motor cortex",
abstract = "Several neural decoding algorithms have successfully converted brain signals into commands to control a computer cursor and prosthetic devices. A majority of decoding methods, such as population vector algorithms (PVA), optimal linear estimators (OLE), and neural networks (NN), are effective in predicting movement kinematics, including movement direction, speed and trajectory but usually require a large number of neurons to achieve desirable performance. This study proposed a novel decoding algorithm even with signals obtained from a smaller numbers of neurons. We adopted sliced inverse regression (SIR) to predict forelimb movement from single-unit activities recorded in the rat primary motor (M1) cortex in a water-reward lever-pressing task. SIR performed weighted principal component analysis (PCA) to achieve effective dimension reduction for nonlinear regression. To demonstrate the decoding performance, SIR was compared to PVA, OLE, and NN. Furthermore, PCA and sequential feature selection (SFS) which are popular feature selection techniques were implemented for comparison of feature selection effectiveness. Among SIR, PVA, OLE, PCA, SFS, and NN decoding methods, the trajectories predicted by SIR (with a root mean square error, RMSE, of 8.47 ± 1.32 mm) was closer to the actual trajectories compared with those predicted by PVA (30.41 ± 11.73 mm), OLE (20.17 ± 6.43 mm), PCA (19.13 ± 0.75 mm), SFS (22.75 ± 2.01 mm), and NN (16.75 ± 2.02 mm). The superiority of SIR was most obvious when the sample size of neurons was small. We concluded that SIR sorted the input data to obtain the effective transform matrices for movement prediction, making it a robust decoding method for conditions with sparse neuronal information.",
keywords = "Forelimb movement prediction, Neural decoding, Neural networks (NN), Principle component analysis (PCA), Sliced inverse regression (SIR)",
author = "Shih-Hung Yang and Chen, {You Yin} and Lin, {Sheng Huang} and Liao, {Lun De} and Lu, {Henry Horng Shing} and Wang, {Ching Fu} and Chen, {Po Chuan} and Lo, {Yu Chun} and Phan, {Thanh Dat} and Chao, {Hsiang Ya} and Lin, {Ching Hui} and Lai, {Hsin Yi} and Huang, {Wei Chen}",
year = "2016",
doi = "10.3389/fnins.2016.00556",
language = "English",
volume = "10",
journal = "Frontiers in Neuroscience",
issn = "1662-4548",
publisher = "Frontiers Research Foundation",
number = "DEC",

}

TY - JOUR

T1 - A sliced inverse regression (SIR) decoding the forelimb movement from neuronal spikes in the rat motor cortex

AU - Yang, Shih-Hung

AU - Chen, You Yin

AU - Lin, Sheng Huang

AU - Liao, Lun De

AU - Lu, Henry Horng Shing

AU - Wang, Ching Fu

AU - Chen, Po Chuan

AU - Lo, Yu Chun

AU - Phan, Thanh Dat

AU - Chao, Hsiang Ya

AU - Lin, Ching Hui

AU - Lai, Hsin Yi

AU - Huang, Wei Chen

PY - 2016

Y1 - 2016

N2 - Several neural decoding algorithms have successfully converted brain signals into commands to control a computer cursor and prosthetic devices. A majority of decoding methods, such as population vector algorithms (PVA), optimal linear estimators (OLE), and neural networks (NN), are effective in predicting movement kinematics, including movement direction, speed and trajectory but usually require a large number of neurons to achieve desirable performance. This study proposed a novel decoding algorithm even with signals obtained from a smaller numbers of neurons. We adopted sliced inverse regression (SIR) to predict forelimb movement from single-unit activities recorded in the rat primary motor (M1) cortex in a water-reward lever-pressing task. SIR performed weighted principal component analysis (PCA) to achieve effective dimension reduction for nonlinear regression. To demonstrate the decoding performance, SIR was compared to PVA, OLE, and NN. Furthermore, PCA and sequential feature selection (SFS) which are popular feature selection techniques were implemented for comparison of feature selection effectiveness. Among SIR, PVA, OLE, PCA, SFS, and NN decoding methods, the trajectories predicted by SIR (with a root mean square error, RMSE, of 8.47 ± 1.32 mm) was closer to the actual trajectories compared with those predicted by PVA (30.41 ± 11.73 mm), OLE (20.17 ± 6.43 mm), PCA (19.13 ± 0.75 mm), SFS (22.75 ± 2.01 mm), and NN (16.75 ± 2.02 mm). The superiority of SIR was most obvious when the sample size of neurons was small. We concluded that SIR sorted the input data to obtain the effective transform matrices for movement prediction, making it a robust decoding method for conditions with sparse neuronal information.

AB - Several neural decoding algorithms have successfully converted brain signals into commands to control a computer cursor and prosthetic devices. A majority of decoding methods, such as population vector algorithms (PVA), optimal linear estimators (OLE), and neural networks (NN), are effective in predicting movement kinematics, including movement direction, speed and trajectory but usually require a large number of neurons to achieve desirable performance. This study proposed a novel decoding algorithm even with signals obtained from a smaller numbers of neurons. We adopted sliced inverse regression (SIR) to predict forelimb movement from single-unit activities recorded in the rat primary motor (M1) cortex in a water-reward lever-pressing task. SIR performed weighted principal component analysis (PCA) to achieve effective dimension reduction for nonlinear regression. To demonstrate the decoding performance, SIR was compared to PVA, OLE, and NN. Furthermore, PCA and sequential feature selection (SFS) which are popular feature selection techniques were implemented for comparison of feature selection effectiveness. Among SIR, PVA, OLE, PCA, SFS, and NN decoding methods, the trajectories predicted by SIR (with a root mean square error, RMSE, of 8.47 ± 1.32 mm) was closer to the actual trajectories compared with those predicted by PVA (30.41 ± 11.73 mm), OLE (20.17 ± 6.43 mm), PCA (19.13 ± 0.75 mm), SFS (22.75 ± 2.01 mm), and NN (16.75 ± 2.02 mm). The superiority of SIR was most obvious when the sample size of neurons was small. We concluded that SIR sorted the input data to obtain the effective transform matrices for movement prediction, making it a robust decoding method for conditions with sparse neuronal information.

KW - Forelimb movement prediction

KW - Neural decoding

KW - Neural networks (NN)

KW - Principle component analysis (PCA)

KW - Sliced inverse regression (SIR)

UR - http://www.scopus.com/inward/record.url?scp=85009811788&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85009811788&partnerID=8YFLogxK

U2 - 10.3389/fnins.2016.00556

DO - 10.3389/fnins.2016.00556

M3 - Article

AN - SCOPUS:85009811788

VL - 10

JO - Frontiers in Neuroscience

JF - Frontiers in Neuroscience

SN - 1662-4548

IS - DEC

M1 - 556

ER -