Analysis of ecog features for movement execution using denoising source separation

Aysegul Gunduz, Justin C. Sanchez, Jose C. Principe

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

1 Citation (Scopus)

Abstract

The major challenge in ECoG-based neuroprosthesis is isolating features in a spectrally and spatially broad range of sources essential for modeling motor behavior. In this study, movement-related spectral modulations are resolved using broadband ECoG recordings passed through a filterbank of constant-Q filters. Denoising source separation is a semiblind source separation methodology which extracts hidden structures of interest within the data by exploiting prior knowledge on the observations. Herein, the methodology is utilized to extract sources that modulate within the frequency content of the hand trajectory. High signal acquisition rates (12kHz) allow for analysis of frequencies beyond the fast gamma oscillations which have been thus far discarded as background activity. Exploratory analysis suggests the first components extracted from envelopes of high spectral bands correlate with the hand trajectory and their spatial distribution covers areas of premotor and primary motor cortices.

Original languageEnglish
Title of host publicationProceedings of the 2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008
Pages103-108
Number of pages6
DOIs
StatePublished - Dec 1 2008
Externally publishedYes
Event2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008 - Cancun, Mexico
Duration: Oct 16 2008Oct 19 2008

Other

Other2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008
CountryMexico
CityCancun
Period10/16/0810/19/08

Fingerprint

Source separation
Trajectories
Spatial distribution
Modulation

ASJC Scopus subject areas

  • Artificial Intelligence
  • Software
  • Electrical and Electronic Engineering

Cite this

Gunduz, A., Sanchez, J. C., & Principe, J. C. (2008). Analysis of ecog features for movement execution using denoising source separation. In Proceedings of the 2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008 (pp. 103-108). [4685463] https://doi.org/10.1109/MLSP.2008.4685463

Analysis of ecog features for movement execution using denoising source separation. / Gunduz, Aysegul; Sanchez, Justin C.; Principe, Jose C.

Proceedings of the 2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008. 2008. p. 103-108 4685463.

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

Gunduz, A, Sanchez, JC & Principe, JC 2008, Analysis of ecog features for movement execution using denoising source separation. in Proceedings of the 2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008., 4685463, pp. 103-108, 2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008, Cancun, Mexico, 10/16/08. https://doi.org/10.1109/MLSP.2008.4685463
Gunduz A, Sanchez JC, Principe JC. Analysis of ecog features for movement execution using denoising source separation. In Proceedings of the 2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008. 2008. p. 103-108. 4685463 https://doi.org/10.1109/MLSP.2008.4685463
Gunduz, Aysegul ; Sanchez, Justin C. ; Principe, Jose C. / Analysis of ecog features for movement execution using denoising source separation. Proceedings of the 2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008. 2008. pp. 103-108
@inproceedings{b97b37bffe754da1a8369a40bd744f6c,
title = "Analysis of ecog features for movement execution using denoising source separation",
abstract = "The major challenge in ECoG-based neuroprosthesis is isolating features in a spectrally and spatially broad range of sources essential for modeling motor behavior. In this study, movement-related spectral modulations are resolved using broadband ECoG recordings passed through a filterbank of constant-Q filters. Denoising source separation is a semiblind source separation methodology which extracts hidden structures of interest within the data by exploiting prior knowledge on the observations. Herein, the methodology is utilized to extract sources that modulate within the frequency content of the hand trajectory. High signal acquisition rates (12kHz) allow for analysis of frequencies beyond the fast gamma oscillations which have been thus far discarded as background activity. Exploratory analysis suggests the first components extracted from envelopes of high spectral bands correlate with the hand trajectory and their spatial distribution covers areas of premotor and primary motor cortices.",
author = "Aysegul Gunduz and Sanchez, {Justin C.} and Principe, {Jose C.}",
year = "2008",
month = "12",
day = "1",
doi = "10.1109/MLSP.2008.4685463",
language = "English",
isbn = "9781424423767",
pages = "103--108",
booktitle = "Proceedings of the 2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008",

}

TY - GEN

T1 - Analysis of ecog features for movement execution using denoising source separation

AU - Gunduz, Aysegul

AU - Sanchez, Justin C.

AU - Principe, Jose C.

PY - 2008/12/1

Y1 - 2008/12/1

N2 - The major challenge in ECoG-based neuroprosthesis is isolating features in a spectrally and spatially broad range of sources essential for modeling motor behavior. In this study, movement-related spectral modulations are resolved using broadband ECoG recordings passed through a filterbank of constant-Q filters. Denoising source separation is a semiblind source separation methodology which extracts hidden structures of interest within the data by exploiting prior knowledge on the observations. Herein, the methodology is utilized to extract sources that modulate within the frequency content of the hand trajectory. High signal acquisition rates (12kHz) allow for analysis of frequencies beyond the fast gamma oscillations which have been thus far discarded as background activity. Exploratory analysis suggests the first components extracted from envelopes of high spectral bands correlate with the hand trajectory and their spatial distribution covers areas of premotor and primary motor cortices.

AB - The major challenge in ECoG-based neuroprosthesis is isolating features in a spectrally and spatially broad range of sources essential for modeling motor behavior. In this study, movement-related spectral modulations are resolved using broadband ECoG recordings passed through a filterbank of constant-Q filters. Denoising source separation is a semiblind source separation methodology which extracts hidden structures of interest within the data by exploiting prior knowledge on the observations. Herein, the methodology is utilized to extract sources that modulate within the frequency content of the hand trajectory. High signal acquisition rates (12kHz) allow for analysis of frequencies beyond the fast gamma oscillations which have been thus far discarded as background activity. Exploratory analysis suggests the first components extracted from envelopes of high spectral bands correlate with the hand trajectory and their spatial distribution covers areas of premotor and primary motor cortices.

UR - http://www.scopus.com/inward/record.url?scp=58049159579&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=58049159579&partnerID=8YFLogxK

U2 - 10.1109/MLSP.2008.4685463

DO - 10.1109/MLSP.2008.4685463

M3 - Conference contribution

SN - 9781424423767

SP - 103

EP - 108

BT - Proceedings of the 2008 IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008

ER -