Statistical analysis for multiplicatively modulated nonlinear autoregressive model and its applications to electrophysiological signal analysis in humans

Hiroko Solvang, Masanobu Taniguchi, Manabu Honda

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

Modulating the dynamics of a nonlinear autoregressive model with a radial basis function (RBF) of exogenous variables is known to reduce the prediction error. Here, RBF is a function that decays to zero exponentially if the deviation between the exogenous variables and a center location becomes large. This paper introduces a class of RBF-based multiplicatively modulated nonlinear autoregressive (mmNAR) models. First, we establish the local asymptotic normality (LAN) for vector conditional heteroscedastic autoregressive nonlinear (CHARN) models, which include the mmNAR and many other well-known time-series models as special cases. Asymptotic optimality for estimation and testing is described in terms of LAN properties. The mmNAR model indicates goodness-of-fit for surface electromyograms (EMG) using electrocorticograms (ECoG) as the exogenous variables. Concretely, it is found that the negative potential of the motor cortex forces change in the frequency of EMG, which is reasonable from a physiological point of view. The proposed mmNAR model fitting is both useful and efficient as a signal-processing technique for extracting information on the action potential, which is associated with the postsynaptic potential.

Original languageEnglish (US)
Pages (from-to)3414-3425
Number of pages12
JournalIEEE Transactions on Signal Processing
Volume54
Issue number9
DOIs
StatePublished - Sep 1 2006

Fingerprint

Signal analysis
Statistical methods
Time series
Signal processing
Testing

Keywords

  • CHARN model
  • Electrocorticogram
  • Electromyogram
  • Local asymptotic normality
  • Radial basis function
  • mmNAR model

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Statistical analysis for multiplicatively modulated nonlinear autoregressive model and its applications to electrophysiological signal analysis in humans. / Solvang, Hiroko; Taniguchi, Masanobu; Honda, Manabu.

In: IEEE Transactions on Signal Processing, Vol. 54, No. 9, 01.09.2006, p. 3414-3425.

Research output: Contribution to journalArticle

@article{b3fcb4c5608f4434b53972e2d60a5d36,
title = "Statistical analysis for multiplicatively modulated nonlinear autoregressive model and its applications to electrophysiological signal analysis in humans",
abstract = "Modulating the dynamics of a nonlinear autoregressive model with a radial basis function (RBF) of exogenous variables is known to reduce the prediction error. Here, RBF is a function that decays to zero exponentially if the deviation between the exogenous variables and a center location becomes large. This paper introduces a class of RBF-based multiplicatively modulated nonlinear autoregressive (mmNAR) models. First, we establish the local asymptotic normality (LAN) for vector conditional heteroscedastic autoregressive nonlinear (CHARN) models, which include the mmNAR and many other well-known time-series models as special cases. Asymptotic optimality for estimation and testing is described in terms of LAN properties. The mmNAR model indicates goodness-of-fit for surface electromyograms (EMG) using electrocorticograms (ECoG) as the exogenous variables. Concretely, it is found that the negative potential of the motor cortex forces change in the frequency of EMG, which is reasonable from a physiological point of view. The proposed mmNAR model fitting is both useful and efficient as a signal-processing technique for extracting information on the action potential, which is associated with the postsynaptic potential.",
keywords = "CHARN model, Electrocorticogram, Electromyogram, Local asymptotic normality, Radial basis function, mmNAR model",
author = "Hiroko Solvang and Masanobu Taniguchi and Manabu Honda",
year = "2006",
month = "9",
day = "1",
doi = "10.1109/TSP.2006.877663",
language = "English (US)",
volume = "54",
pages = "3414--3425",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "9",

}

TY - JOUR

T1 - Statistical analysis for multiplicatively modulated nonlinear autoregressive model and its applications to electrophysiological signal analysis in humans

AU - Solvang, Hiroko

AU - Taniguchi, Masanobu

AU - Honda, Manabu

PY - 2006/9/1

Y1 - 2006/9/1

N2 - Modulating the dynamics of a nonlinear autoregressive model with a radial basis function (RBF) of exogenous variables is known to reduce the prediction error. Here, RBF is a function that decays to zero exponentially if the deviation between the exogenous variables and a center location becomes large. This paper introduces a class of RBF-based multiplicatively modulated nonlinear autoregressive (mmNAR) models. First, we establish the local asymptotic normality (LAN) for vector conditional heteroscedastic autoregressive nonlinear (CHARN) models, which include the mmNAR and many other well-known time-series models as special cases. Asymptotic optimality for estimation and testing is described in terms of LAN properties. The mmNAR model indicates goodness-of-fit for surface electromyograms (EMG) using electrocorticograms (ECoG) as the exogenous variables. Concretely, it is found that the negative potential of the motor cortex forces change in the frequency of EMG, which is reasonable from a physiological point of view. The proposed mmNAR model fitting is both useful and efficient as a signal-processing technique for extracting information on the action potential, which is associated with the postsynaptic potential.

AB - Modulating the dynamics of a nonlinear autoregressive model with a radial basis function (RBF) of exogenous variables is known to reduce the prediction error. Here, RBF is a function that decays to zero exponentially if the deviation between the exogenous variables and a center location becomes large. This paper introduces a class of RBF-based multiplicatively modulated nonlinear autoregressive (mmNAR) models. First, we establish the local asymptotic normality (LAN) for vector conditional heteroscedastic autoregressive nonlinear (CHARN) models, which include the mmNAR and many other well-known time-series models as special cases. Asymptotic optimality for estimation and testing is described in terms of LAN properties. The mmNAR model indicates goodness-of-fit for surface electromyograms (EMG) using electrocorticograms (ECoG) as the exogenous variables. Concretely, it is found that the negative potential of the motor cortex forces change in the frequency of EMG, which is reasonable from a physiological point of view. The proposed mmNAR model fitting is both useful and efficient as a signal-processing technique for extracting information on the action potential, which is associated with the postsynaptic potential.

KW - CHARN model

KW - Electrocorticogram

KW - Electromyogram

KW - Local asymptotic normality

KW - Radial basis function

KW - mmNAR model

UR - http://www.scopus.com/inward/record.url?scp=33947171438&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33947171438&partnerID=8YFLogxK

U2 - 10.1109/TSP.2006.877663

DO - 10.1109/TSP.2006.877663

M3 - Article

AN - SCOPUS:33947171438

VL - 54

SP - 3414

EP - 3425

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 9

ER -