Decoding hand trajectories from ECoG recordings via kernel least-mean-square algorithm

Aysegul Gunduz, Jung Phil Kwon, Justin C. Sanchez, Jose C. Principe

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

5 Citations (Scopus)

Abstract

Prediction of two dimensional hand trajectories from cortical surface recordings entails finding a functional mapping from spectral modulations in multidimensional channels to instantaneous hand positions. Such studies thus far have been conducted through linear adaptive filters, even though, the functional mapping from the cortical activity to behavior might be nonlinear. Herein, we employ a nonlinear adaptive filter, kernel least mean square (KLMS), which nonlinearly map inputs to a higher dimensional feature space in which inner products can be efficiently computed. The methodology is a simple and effective nonlinear extension of the least mean square (LMS) algorithm. Preliminary results show significant improvements in mean squared error (MSE) values of reconstructed trajectories compared to linear methods (LMS) at a confidence level of 95% in the axis of highest excursion.

Original languageEnglish
Title of host publication2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09
Pages267-270
Number of pages4
DOIs
StatePublished - Oct 27 2009
Externally publishedYes
Event2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09 - Antalya, Turkey
Duration: Apr 29 2009May 2 2009

Other

Other2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09
CountryTurkey
CityAntalya
Period4/29/095/2/09

Fingerprint

Adaptive filters
Least-Squares Analysis
Decoding
Hand
Trajectories
Modulation

ASJC Scopus subject areas

  • Biomedical Engineering
  • Clinical Neurology
  • Neuroscience(all)

Cite this

Gunduz, A., Kwon, J. P., Sanchez, J. C., & Principe, J. C. (2009). Decoding hand trajectories from ECoG recordings via kernel least-mean-square algorithm. In 2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09 (pp. 267-270). [5109284] https://doi.org/10.1109/NER.2009.5109284

Decoding hand trajectories from ECoG recordings via kernel least-mean-square algorithm. / Gunduz, Aysegul; Kwon, Jung Phil; Sanchez, Justin C.; Principe, Jose C.

2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09. 2009. p. 267-270 5109284.

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

Gunduz, A, Kwon, JP, Sanchez, JC & Principe, JC 2009, Decoding hand trajectories from ECoG recordings via kernel least-mean-square algorithm. in 2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09., 5109284, pp. 267-270, 2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09, Antalya, Turkey, 4/29/09. https://doi.org/10.1109/NER.2009.5109284
Gunduz A, Kwon JP, Sanchez JC, Principe JC. Decoding hand trajectories from ECoG recordings via kernel least-mean-square algorithm. In 2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09. 2009. p. 267-270. 5109284 https://doi.org/10.1109/NER.2009.5109284
Gunduz, Aysegul ; Kwon, Jung Phil ; Sanchez, Justin C. ; Principe, Jose C. / Decoding hand trajectories from ECoG recordings via kernel least-mean-square algorithm. 2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09. 2009. pp. 267-270
@inproceedings{1a492e099ab94e28afa4c65c0e331391,
title = "Decoding hand trajectories from ECoG recordings via kernel least-mean-square algorithm",
abstract = "Prediction of two dimensional hand trajectories from cortical surface recordings entails finding a functional mapping from spectral modulations in multidimensional channels to instantaneous hand positions. Such studies thus far have been conducted through linear adaptive filters, even though, the functional mapping from the cortical activity to behavior might be nonlinear. Herein, we employ a nonlinear adaptive filter, kernel least mean square (KLMS), which nonlinearly map inputs to a higher dimensional feature space in which inner products can be efficiently computed. The methodology is a simple and effective nonlinear extension of the least mean square (LMS) algorithm. Preliminary results show significant improvements in mean squared error (MSE) values of reconstructed trajectories compared to linear methods (LMS) at a confidence level of 95{\%} in the axis of highest excursion.",
author = "Aysegul Gunduz and Kwon, {Jung Phil} and Sanchez, {Justin C.} and Principe, {Jose C.}",
year = "2009",
month = "10",
day = "27",
doi = "10.1109/NER.2009.5109284",
language = "English",
isbn = "9781424420735",
pages = "267--270",
booktitle = "2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09",

}

TY - GEN

T1 - Decoding hand trajectories from ECoG recordings via kernel least-mean-square algorithm

AU - Gunduz, Aysegul

AU - Kwon, Jung Phil

AU - Sanchez, Justin C.

AU - Principe, Jose C.

PY - 2009/10/27

Y1 - 2009/10/27

N2 - Prediction of two dimensional hand trajectories from cortical surface recordings entails finding a functional mapping from spectral modulations in multidimensional channels to instantaneous hand positions. Such studies thus far have been conducted through linear adaptive filters, even though, the functional mapping from the cortical activity to behavior might be nonlinear. Herein, we employ a nonlinear adaptive filter, kernel least mean square (KLMS), which nonlinearly map inputs to a higher dimensional feature space in which inner products can be efficiently computed. The methodology is a simple and effective nonlinear extension of the least mean square (LMS) algorithm. Preliminary results show significant improvements in mean squared error (MSE) values of reconstructed trajectories compared to linear methods (LMS) at a confidence level of 95% in the axis of highest excursion.

AB - Prediction of two dimensional hand trajectories from cortical surface recordings entails finding a functional mapping from spectral modulations in multidimensional channels to instantaneous hand positions. Such studies thus far have been conducted through linear adaptive filters, even though, the functional mapping from the cortical activity to behavior might be nonlinear. Herein, we employ a nonlinear adaptive filter, kernel least mean square (KLMS), which nonlinearly map inputs to a higher dimensional feature space in which inner products can be efficiently computed. The methodology is a simple and effective nonlinear extension of the least mean square (LMS) algorithm. Preliminary results show significant improvements in mean squared error (MSE) values of reconstructed trajectories compared to linear methods (LMS) at a confidence level of 95% in the axis of highest excursion.

UR - http://www.scopus.com/inward/record.url?scp=70350220605&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70350220605&partnerID=8YFLogxK

U2 - 10.1109/NER.2009.5109284

DO - 10.1109/NER.2009.5109284

M3 - Conference contribution

AN - SCOPUS:70350220605

SN - 9781424420735

SP - 267

EP - 270

BT - 2009 4th International IEEE/EMBS Conference on Neural Engineering, NER '09

ER -