Including planning activity in feature space distributes activation over a broader neuron population

Girish Singhal, Soumyadipta Acharya, Natan Davidovics, He Jiping, Nitish V Thakor

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In neuroprosthetic systems, decoding based on a sparse population of task-related neurons is impractical because micro-electrode arrays often drift gradually in the cortex. Since the neuronal population being recorded from is dynamic, it is favorable to have a larger number of neurons containing information relevant to movement decoding and to decrease the relative importance of individual neurons. We have shown that a feature space comprised of neural firing rates from planning as well as movement periods exists in a broader distribution of neurons, as compared to a feature space that is derived from the movement period alone. For this study, spike train data from 297 neurons located in M1 and PM areas was analyzed to validate the hypothesis. The data was from a rhesus monkey performing reach to grasp task with measured wrist supination/pronation. Artificial neural networks were used to model encoding of wrist angle, and a sensitivity analysis was performed to attribute the relative importance of the input neurons. A. classifier trained on only the least important neurons, as determined by their relative contribution to the decoded variable, had an average 20% better decoding accuracy when the new method of feature selection was used. This indicates that there is valuable information content within the distributed neuronal population.

Original languageEnglish (US)
Title of host publicationAnnual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings
Pages5349-5352
Number of pages4
DOIs
StatePublished - 2007
Event29th Annual International Conference of IEEE-EMBS, Engineering in Medicine and Biology Society, EMBC'07 - Lyon, France
Duration: Aug 23 2007Aug 26 2007

Other

Other29th Annual International Conference of IEEE-EMBS, Engineering in Medicine and Biology Society, EMBC'07
CountryFrance
CityLyon
Period8/23/078/26/07

Fingerprint

Neurons
Chemical activation
Planning
Decoding
Sensitivity analysis
Feature extraction
Classifiers
Neural networks
Electrodes

Keywords

  • Brain-machine interface
  • Broad tuning curves
  • Neural decoding
  • Neural prosthesis
  • Sensitivity analysis

ASJC Scopus subject areas

  • Biomedical Engineering

Cite this

Singhal, G., Acharya, S., Davidovics, N., Jiping, H., & Thakor, N. V. (2007). Including planning activity in feature space distributes activation over a broader neuron population. In Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings (pp. 5349-5352). [4353550] https://doi.org/10.1109/IEMBS.2007.4353550

Including planning activity in feature space distributes activation over a broader neuron population. / Singhal, Girish; Acharya, Soumyadipta; Davidovics, Natan; Jiping, He; Thakor, Nitish V.

Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings. 2007. p. 5349-5352 4353550.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Singhal, G, Acharya, S, Davidovics, N, Jiping, H & Thakor, NV 2007, Including planning activity in feature space distributes activation over a broader neuron population. in Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings., 4353550, pp. 5349-5352, 29th Annual International Conference of IEEE-EMBS, Engineering in Medicine and Biology Society, EMBC'07, Lyon, France, 8/23/07. https://doi.org/10.1109/IEMBS.2007.4353550
Singhal G, Acharya S, Davidovics N, Jiping H, Thakor NV. Including planning activity in feature space distributes activation over a broader neuron population. In Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings. 2007. p. 5349-5352. 4353550 https://doi.org/10.1109/IEMBS.2007.4353550
Singhal, Girish ; Acharya, Soumyadipta ; Davidovics, Natan ; Jiping, He ; Thakor, Nitish V. / Including planning activity in feature space distributes activation over a broader neuron population. Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings. 2007. pp. 5349-5352
@inproceedings{08a88dd3ebe349ea8b04a16eea67c13f,
title = "Including planning activity in feature space distributes activation over a broader neuron population",
abstract = "In neuroprosthetic systems, decoding based on a sparse population of task-related neurons is impractical because micro-electrode arrays often drift gradually in the cortex. Since the neuronal population being recorded from is dynamic, it is favorable to have a larger number of neurons containing information relevant to movement decoding and to decrease the relative importance of individual neurons. We have shown that a feature space comprised of neural firing rates from planning as well as movement periods exists in a broader distribution of neurons, as compared to a feature space that is derived from the movement period alone. For this study, spike train data from 297 neurons located in M1 and PM areas was analyzed to validate the hypothesis. The data was from a rhesus monkey performing reach to grasp task with measured wrist supination/pronation. Artificial neural networks were used to model encoding of wrist angle, and a sensitivity analysis was performed to attribute the relative importance of the input neurons. A. classifier trained on only the least important neurons, as determined by their relative contribution to the decoded variable, had an average 20{\%} better decoding accuracy when the new method of feature selection was used. This indicates that there is valuable information content within the distributed neuronal population.",
keywords = "Brain-machine interface, Broad tuning curves, Neural decoding, Neural prosthesis, Sensitivity analysis",
author = "Girish Singhal and Soumyadipta Acharya and Natan Davidovics and He Jiping and Thakor, {Nitish V}",
year = "2007",
doi = "10.1109/IEMBS.2007.4353550",
language = "English (US)",
isbn = "1424407885",
pages = "5349--5352",
booktitle = "Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings",

}

TY - GEN

T1 - Including planning activity in feature space distributes activation over a broader neuron population

AU - Singhal, Girish

AU - Acharya, Soumyadipta

AU - Davidovics, Natan

AU - Jiping, He

AU - Thakor, Nitish V

PY - 2007

Y1 - 2007

N2 - In neuroprosthetic systems, decoding based on a sparse population of task-related neurons is impractical because micro-electrode arrays often drift gradually in the cortex. Since the neuronal population being recorded from is dynamic, it is favorable to have a larger number of neurons containing information relevant to movement decoding and to decrease the relative importance of individual neurons. We have shown that a feature space comprised of neural firing rates from planning as well as movement periods exists in a broader distribution of neurons, as compared to a feature space that is derived from the movement period alone. For this study, spike train data from 297 neurons located in M1 and PM areas was analyzed to validate the hypothesis. The data was from a rhesus monkey performing reach to grasp task with measured wrist supination/pronation. Artificial neural networks were used to model encoding of wrist angle, and a sensitivity analysis was performed to attribute the relative importance of the input neurons. A. classifier trained on only the least important neurons, as determined by their relative contribution to the decoded variable, had an average 20% better decoding accuracy when the new method of feature selection was used. This indicates that there is valuable information content within the distributed neuronal population.

AB - In neuroprosthetic systems, decoding based on a sparse population of task-related neurons is impractical because micro-electrode arrays often drift gradually in the cortex. Since the neuronal population being recorded from is dynamic, it is favorable to have a larger number of neurons containing information relevant to movement decoding and to decrease the relative importance of individual neurons. We have shown that a feature space comprised of neural firing rates from planning as well as movement periods exists in a broader distribution of neurons, as compared to a feature space that is derived from the movement period alone. For this study, spike train data from 297 neurons located in M1 and PM areas was analyzed to validate the hypothesis. The data was from a rhesus monkey performing reach to grasp task with measured wrist supination/pronation. Artificial neural networks were used to model encoding of wrist angle, and a sensitivity analysis was performed to attribute the relative importance of the input neurons. A. classifier trained on only the least important neurons, as determined by their relative contribution to the decoded variable, had an average 20% better decoding accuracy when the new method of feature selection was used. This indicates that there is valuable information content within the distributed neuronal population.

KW - Brain-machine interface

KW - Broad tuning curves

KW - Neural decoding

KW - Neural prosthesis

KW - Sensitivity analysis

UR - http://www.scopus.com/inward/record.url?scp=57649181793&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=57649181793&partnerID=8YFLogxK

U2 - 10.1109/IEMBS.2007.4353550

DO - 10.1109/IEMBS.2007.4353550

M3 - Conference contribution

C2 - 18003216

AN - SCOPUS:57649181793

SN - 1424407885

SN - 9781424407880

SP - 5349

EP - 5352

BT - Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings

ER -