On the performance of autocorrelation estimation algorithms for fMRI analysis

Brian Lenoski, Lislie C. Baxter, Lina Karam, José Maisog, Josef Debbins

Research output: Contribution to journalArticle

31 Citations (Scopus)

Abstract

Pre-whitening of fMRI time-series is commonly performed to address temporal autocorrelations. The pre-whitening procedure requires knowledge of the spatially dependent autocorrelations, which in turn must be estimated from the observed data. The accuracy of the autocorrelation estimation algorithm is important because biased autocorrelation estimates result in biased test statistics, thereby increasing the expected false-positive and/or false-negative rates. Thus, a methodology for testing the accuracy of autocorrelation estimates and for assessing the performance of today's state-of-the-art autocorrelation estimation algorithms is needed. To address these problems, we propose an evaluation framework that tests for significant autocorrelation bias in the model residuals of a general linear model analysis. We apply the proposed testing framework to 18 pre-surgical fMRI mapping datasets from ten patients and compare the performance of popular fMRI autocorrelation estimation algorithms. We also identify five consistent spectral patterns representative of the encountered autocorrelation structures and show that they are well described by a second-order/two-pole model. We subsequently show that a nonregularized, second-order autoregressive model, AR(2), is sufficient for capturing the range of temporal autocorrelations found in the considered fMRI datasets. Finally, we explore the bias versus predictability tradeoff associated with regularization of the autocorrelation coefficients. We find that the increased bias from regularization outweighs any gains in predictability. Based on the obtained results, we expect that a second-order, nonregularized AR algorithm will provide the best performance in terms of producing white residuals and achieving the best possible tradeoff between maximizing predictability and minimizing bias for most fMRI datasets.

Original languageEnglish (US)
Pages (from-to)828-838
Number of pages11
JournalIEEE Journal on Selected Topics in Signal Processing
Volume2
Issue number6
DOIs
StatePublished - 2008

Fingerprint

Autocorrelation
Magnetic Resonance Imaging
Testing
Time series
Poles
Statistics

Keywords

  • Algorithm design and analysis
  • Autocorrelation
  • Bias
  • Brain mapping
  • Estimation
  • fMRI
  • Magnetic resonance imaging
  • Pre-whitening
  • Regularization

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Signal Processing

Cite this

On the performance of autocorrelation estimation algorithms for fMRI analysis. / Lenoski, Brian; Baxter, Lislie C.; Karam, Lina; Maisog, José; Debbins, Josef.

In: IEEE Journal on Selected Topics in Signal Processing, Vol. 2, No. 6, 2008, p. 828-838.

Research output: Contribution to journalArticle

Lenoski, Brian ; Baxter, Lislie C. ; Karam, Lina ; Maisog, José ; Debbins, Josef. / On the performance of autocorrelation estimation algorithms for fMRI analysis. In: IEEE Journal on Selected Topics in Signal Processing. 2008 ; Vol. 2, No. 6. pp. 828-838.
@article{448855d04ebb45bba5de425dfb8cddb3,
title = "On the performance of autocorrelation estimation algorithms for fMRI analysis",
abstract = "Pre-whitening of fMRI time-series is commonly performed to address temporal autocorrelations. The pre-whitening procedure requires knowledge of the spatially dependent autocorrelations, which in turn must be estimated from the observed data. The accuracy of the autocorrelation estimation algorithm is important because biased autocorrelation estimates result in biased test statistics, thereby increasing the expected false-positive and/or false-negative rates. Thus, a methodology for testing the accuracy of autocorrelation estimates and for assessing the performance of today's state-of-the-art autocorrelation estimation algorithms is needed. To address these problems, we propose an evaluation framework that tests for significant autocorrelation bias in the model residuals of a general linear model analysis. We apply the proposed testing framework to 18 pre-surgical fMRI mapping datasets from ten patients and compare the performance of popular fMRI autocorrelation estimation algorithms. We also identify five consistent spectral patterns representative of the encountered autocorrelation structures and show that they are well described by a second-order/two-pole model. We subsequently show that a nonregularized, second-order autoregressive model, AR(2), is sufficient for capturing the range of temporal autocorrelations found in the considered fMRI datasets. Finally, we explore the bias versus predictability tradeoff associated with regularization of the autocorrelation coefficients. We find that the increased bias from regularization outweighs any gains in predictability. Based on the obtained results, we expect that a second-order, nonregularized AR algorithm will provide the best performance in terms of producing white residuals and achieving the best possible tradeoff between maximizing predictability and minimizing bias for most fMRI datasets.",
keywords = "Algorithm design and analysis, Autocorrelation, Bias, Brain mapping, Estimation, fMRI, Magnetic resonance imaging, Pre-whitening, Regularization",
author = "Brian Lenoski and Baxter, {Lislie C.} and Lina Karam and Jos{\'e} Maisog and Josef Debbins",
year = "2008",
doi = "10.1109/JSTSP.2008.2007819",
language = "English (US)",
volume = "2",
pages = "828--838",
journal = "IEEE Journal on Selected Topics in Signal Processing",
issn = "1932-4553",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "6",

}

TY - JOUR

T1 - On the performance of autocorrelation estimation algorithms for fMRI analysis

AU - Lenoski, Brian

AU - Baxter, Lislie C.

AU - Karam, Lina

AU - Maisog, José

AU - Debbins, Josef

PY - 2008

Y1 - 2008

N2 - Pre-whitening of fMRI time-series is commonly performed to address temporal autocorrelations. The pre-whitening procedure requires knowledge of the spatially dependent autocorrelations, which in turn must be estimated from the observed data. The accuracy of the autocorrelation estimation algorithm is important because biased autocorrelation estimates result in biased test statistics, thereby increasing the expected false-positive and/or false-negative rates. Thus, a methodology for testing the accuracy of autocorrelation estimates and for assessing the performance of today's state-of-the-art autocorrelation estimation algorithms is needed. To address these problems, we propose an evaluation framework that tests for significant autocorrelation bias in the model residuals of a general linear model analysis. We apply the proposed testing framework to 18 pre-surgical fMRI mapping datasets from ten patients and compare the performance of popular fMRI autocorrelation estimation algorithms. We also identify five consistent spectral patterns representative of the encountered autocorrelation structures and show that they are well described by a second-order/two-pole model. We subsequently show that a nonregularized, second-order autoregressive model, AR(2), is sufficient for capturing the range of temporal autocorrelations found in the considered fMRI datasets. Finally, we explore the bias versus predictability tradeoff associated with regularization of the autocorrelation coefficients. We find that the increased bias from regularization outweighs any gains in predictability. Based on the obtained results, we expect that a second-order, nonregularized AR algorithm will provide the best performance in terms of producing white residuals and achieving the best possible tradeoff between maximizing predictability and minimizing bias for most fMRI datasets.

AB - Pre-whitening of fMRI time-series is commonly performed to address temporal autocorrelations. The pre-whitening procedure requires knowledge of the spatially dependent autocorrelations, which in turn must be estimated from the observed data. The accuracy of the autocorrelation estimation algorithm is important because biased autocorrelation estimates result in biased test statistics, thereby increasing the expected false-positive and/or false-negative rates. Thus, a methodology for testing the accuracy of autocorrelation estimates and for assessing the performance of today's state-of-the-art autocorrelation estimation algorithms is needed. To address these problems, we propose an evaluation framework that tests for significant autocorrelation bias in the model residuals of a general linear model analysis. We apply the proposed testing framework to 18 pre-surgical fMRI mapping datasets from ten patients and compare the performance of popular fMRI autocorrelation estimation algorithms. We also identify five consistent spectral patterns representative of the encountered autocorrelation structures and show that they are well described by a second-order/two-pole model. We subsequently show that a nonregularized, second-order autoregressive model, AR(2), is sufficient for capturing the range of temporal autocorrelations found in the considered fMRI datasets. Finally, we explore the bias versus predictability tradeoff associated with regularization of the autocorrelation coefficients. We find that the increased bias from regularization outweighs any gains in predictability. Based on the obtained results, we expect that a second-order, nonregularized AR algorithm will provide the best performance in terms of producing white residuals and achieving the best possible tradeoff between maximizing predictability and minimizing bias for most fMRI datasets.

KW - Algorithm design and analysis

KW - Autocorrelation

KW - Bias

KW - Brain mapping

KW - Estimation

KW - fMRI

KW - Magnetic resonance imaging

KW - Pre-whitening

KW - Regularization

UR - http://www.scopus.com/inward/record.url?scp=60549096100&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=60549096100&partnerID=8YFLogxK

U2 - 10.1109/JSTSP.2008.2007819

DO - 10.1109/JSTSP.2008.2007819

M3 - Article

VL - 2

SP - 828

EP - 838

JO - IEEE Journal on Selected Topics in Signal Processing

JF - IEEE Journal on Selected Topics in Signal Processing

SN - 1932-4553

IS - 6

ER -