Optimal spline smoothing of fMRI time series by generalized cross-validation

John D. Carew, Grace Wahba, Xianhong Xie, Erik V. Nordheim, M. Elizabeth Meyerand

Research output: Contribution to journalArticle

27 Citations (Scopus)

Abstract

Linear parametric regression models of fMRI time series have correlated residuals. One approach to address this problem is to condition the autocorrelation structure by temporal smoothing. Smoothing splines with the degree of smoothing selected by generalized cross-validation (GCV-spline) provide a method to find an optimal smoother for an fMRI time series. The purpose of this study was to determine if GCV-spline of fMRI time series yields unbiased variance estimates of linear regression model parameters. GCV-spline was evaluated with a real fMRI data set and bias of the variance estimator was computed for simulated time series with autocorrelation structures derived from fMRI data. This study only considered fMRI experimental designs of boxcar type. The results from the real data suggest that GCV-spline determines appropriate amounts of smoothing. The simulations show that the variance estimates are, on average, unbiased. The unbiased variance estimates come at some cost to signal detection efficiency. This study demonstrates that GCV-spline is an appropriate method for smoothing fMRI time series.

Original languageEnglish (US)
Pages (from-to)950-961
Number of pages12
JournalNeuroImage
Volume18
Issue number4
DOIs
StatePublished - Apr 1 2003
Externally publishedYes

Fingerprint

Magnetic Resonance Imaging
Linear Models
Research Design
Costs and Cost Analysis

Keywords

  • Functional MRI
  • General linear model
  • Generalized cross-validation
  • Smoothing spline
  • Temporal autocorrelation

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Neurology

Cite this

Optimal spline smoothing of fMRI time series by generalized cross-validation. / Carew, John D.; Wahba, Grace; Xie, Xianhong; Nordheim, Erik V.; Meyerand, M. Elizabeth.

In: NeuroImage, Vol. 18, No. 4, 01.04.2003, p. 950-961.

Research output: Contribution to journalArticle

Carew, John D. ; Wahba, Grace ; Xie, Xianhong ; Nordheim, Erik V. ; Meyerand, M. Elizabeth. / Optimal spline smoothing of fMRI time series by generalized cross-validation. In: NeuroImage. 2003 ; Vol. 18, No. 4. pp. 950-961.
@article{d18f4d1a660440e8b913bc7a0856578e,
title = "Optimal spline smoothing of fMRI time series by generalized cross-validation",
abstract = "Linear parametric regression models of fMRI time series have correlated residuals. One approach to address this problem is to condition the autocorrelation structure by temporal smoothing. Smoothing splines with the degree of smoothing selected by generalized cross-validation (GCV-spline) provide a method to find an optimal smoother for an fMRI time series. The purpose of this study was to determine if GCV-spline of fMRI time series yields unbiased variance estimates of linear regression model parameters. GCV-spline was evaluated with a real fMRI data set and bias of the variance estimator was computed for simulated time series with autocorrelation structures derived from fMRI data. This study only considered fMRI experimental designs of boxcar type. The results from the real data suggest that GCV-spline determines appropriate amounts of smoothing. The simulations show that the variance estimates are, on average, unbiased. The unbiased variance estimates come at some cost to signal detection efficiency. This study demonstrates that GCV-spline is an appropriate method for smoothing fMRI time series.",
keywords = "Functional MRI, General linear model, Generalized cross-validation, Smoothing spline, Temporal autocorrelation",
author = "Carew, {John D.} and Grace Wahba and Xianhong Xie and Nordheim, {Erik V.} and Meyerand, {M. Elizabeth}",
year = "2003",
month = "4",
day = "1",
doi = "10.1016/S1053-8119(03)00013-2",
language = "English (US)",
volume = "18",
pages = "950--961",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",
number = "4",

}

TY - JOUR

T1 - Optimal spline smoothing of fMRI time series by generalized cross-validation

AU - Carew, John D.

AU - Wahba, Grace

AU - Xie, Xianhong

AU - Nordheim, Erik V.

AU - Meyerand, M. Elizabeth

PY - 2003/4/1

Y1 - 2003/4/1

N2 - Linear parametric regression models of fMRI time series have correlated residuals. One approach to address this problem is to condition the autocorrelation structure by temporal smoothing. Smoothing splines with the degree of smoothing selected by generalized cross-validation (GCV-spline) provide a method to find an optimal smoother for an fMRI time series. The purpose of this study was to determine if GCV-spline of fMRI time series yields unbiased variance estimates of linear regression model parameters. GCV-spline was evaluated with a real fMRI data set and bias of the variance estimator was computed for simulated time series with autocorrelation structures derived from fMRI data. This study only considered fMRI experimental designs of boxcar type. The results from the real data suggest that GCV-spline determines appropriate amounts of smoothing. The simulations show that the variance estimates are, on average, unbiased. The unbiased variance estimates come at some cost to signal detection efficiency. This study demonstrates that GCV-spline is an appropriate method for smoothing fMRI time series.

AB - Linear parametric regression models of fMRI time series have correlated residuals. One approach to address this problem is to condition the autocorrelation structure by temporal smoothing. Smoothing splines with the degree of smoothing selected by generalized cross-validation (GCV-spline) provide a method to find an optimal smoother for an fMRI time series. The purpose of this study was to determine if GCV-spline of fMRI time series yields unbiased variance estimates of linear regression model parameters. GCV-spline was evaluated with a real fMRI data set and bias of the variance estimator was computed for simulated time series with autocorrelation structures derived from fMRI data. This study only considered fMRI experimental designs of boxcar type. The results from the real data suggest that GCV-spline determines appropriate amounts of smoothing. The simulations show that the variance estimates are, on average, unbiased. The unbiased variance estimates come at some cost to signal detection efficiency. This study demonstrates that GCV-spline is an appropriate method for smoothing fMRI time series.

KW - Functional MRI

KW - General linear model

KW - Generalized cross-validation

KW - Smoothing spline

KW - Temporal autocorrelation

UR - http://www.scopus.com/inward/record.url?scp=0037396577&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0037396577&partnerID=8YFLogxK

U2 - 10.1016/S1053-8119(03)00013-2

DO - 10.1016/S1053-8119(03)00013-2

M3 - Article

VL - 18

SP - 950

EP - 961

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

IS - 4

ER -