Elsevier

NeuroImage

Volume 35, Issue 1, March 2007, Pages 396-405
NeuroImage

A method for using blocked and event-related fMRI data to study “resting state” functional connectivity

https://doi.org/10.1016/j.neuroimage.2006.11.051Get rights and content

Abstract

Resting state functional connectivity MRI (fcMRI) has become a particularly useful tool for studying regional relationships in typical and atypical populations. Because many investigators have already obtained large data sets of task-related fMRI, the ability to use this existing task data for resting state fcMRI is of considerable interest. Two classes of data sets could potentially be modified to emulate resting state data. These data sets include: (1) “interleaved” resting blocks from blocked or mixed blocked/event-related sets, and (2) residual timecourses from event-related sets that lack rest blocks. Using correlation analysis, we compared the functional connectivity of resting epochs taken from a mixed blocked/event-related design fMRI data set and the residuals derived from event-related data with standard continuous resting state data to determine which class of data can best emulate resting state data. We show that, despite some differences, the functional connectivity for the interleaved resting periods taken from blocked designs is both qualitatively and quantitatively very similar to that of “continuous” resting state data. In contrast, despite being qualitatively similar to “continuous” resting state data, residuals derived from event-related design data had several distinct quantitative differences. These results suggest that the interleaved resting state data such as those taken from blocked or mixed blocked/event-related fMRI designs are well-suited for resting state functional connectivity analyses. Although using event-related data residuals for resting state functional connectivity may still be useful, results should be interpreted with care.

Introduction

Functional neuroimaging data have been utilized for over a decade to examine regional interactions in the brain. Such interactions have been termed “functional connectivity” and at the “Functional Brain Connectivity Workshop” in Dusseldorf, Germany have been formally defined as the “temporal correlations between spatially remote neurophysiological events” (Friston et al., 1993a, Lee et al., 2003). Functional connectivity analyses were first performed using Positron Emission Tomography (PET) (Clark et al., 1984, Friston et al., 1993b, Horwitz et al., 1984, Horwitz et al., 1992, Metter et al., 1984a, Metter et al., 1984b, Prohovnik et al., 1980) and have since expanded into several other imaging modalities, for which many novel analysis strategies have been developed (Bellec et al., 2006, Biswal et al., 1995, Bokde et al., 2001, Buchel and Friston, 1997, Bullmore et al., 2000, Goncalves et al., 2001, Hampson et al., 2002, Horwitz, 2003, Horwitz et al., 2005, Lee et al., 2003, Pugh et al., 2000).

In functional magnetic resonance imaging (fMRI), attention has recently focused on determining regional interactions using data from subjects at rest (i.e. “resting state functional connectivity”) (Biswal et al., 1995, Fox et al., 2005, Fransson, 2005, Greicius et al., 2003). Resting state functional connectivity measures low frequency (< 0.08 Hz) blood oxygen level-dependent (BOLD) signal fluctuations between regions occurring at rest (Biswal et al., 1995, Fox et al., 2005, Greicius et al., 2003). These low frequency BOLD fluctuations are presumed to relate to “spontaneous” neural activity (Biswal et al., 1995, Leopold et al., 2003, Nir et al., 2006). By cross-correlating the time series of a particular brain region (seed region) with all other voxels, one can determine which voxels are “functionally connected” with that region.

Resting state functional connectivity measures are of interest for several reasons. First, some consider them to reflect human anatomical connectivity (Koch et al., 2002, Quigley et al., 2003). Second, the reliance of functional connectivity MRI (fcMRI) on resting state data unburdens experimental design, subject compliance, and training demands making it attractive for studies of development and clinical groups (Bokde et al., 2006, Greicius et al., 2004, Rombouts and Scheltens, 2005, Tian et al., 2006, Whalley et al., 2005). Because many investigators have already obtained large task-related data sets of atypical populations, the ability to take advantage of existing task data and extract resting state data is of considerable interest.

Two classes of data sets could potentially be modified to emulate resting state data. These sets include the use of (1) interleaved resting blocks from blocked or mixed blocked/event-related sets, and (2) residual timecourses from event-related sets that lack rest blocks.

These two types of data sets present different sets of concerns about their applicability for fcMRI. Using blocked or mixed blocked/event-related fMRI data sets with relatively short resting periods limits the range of frequencies that can be used to extract the resting correlation information; however, it has been suggested that isolating higher frequencies above 0.1 Hz will produce similar correlation profiles (Salvador et al., 2005), but with slightly lower correlation coefficients than if using frequencies below 0.1 Hz. These results suggest that concatenating the resting epochs of some blocked design data, despite missing the lowest frequency components due to the shorter total sampling period, may yield similar correlation profiles compared to continuous resting state data. Another possible concern is that spontaneous resting state activity may be altered by previous task states. There has been limited work on this issue, but the available data suggest that this may only be a concern on the individual subject level (Waites et al., 2005). Moreover, Hampson et al. (2002), while examining an approach for identifying interregional correlations in resting state data using two independent data sets, despite some differences, qualitatively observed similar correlation patterns between the resting periods of a blocked design data set and continuous resting data for a premotor seed region.

Extracting “resting state” data for functional connectivity analyses from event-related fMRI data presents a different set of problems in that (1) task-induced correlations may contaminate the resting correlations, and (2) constant engagement of a task may alter the underlying spontaneous BOLD fluctuations (Fransson, 2006). Work by Fox et al. (2006b) and Arfanakis et al. (2000) have demonstrated that much of the variance observed in trial-to-trial task-evoked activity can be accounted for by the underlying spontaneous activity. Fox and colleagues suggest that there is a linear addition of task-related activity on top of persistent resting spontaneous activity. The implication here is that constant task engagement would not affect the underlying spontaneous BOLD activity. Results thus far supporting the linear superposition premise have been limited to primary sensory and motor regions and specific task conditions (Arfanakis et al., 2000, Fox et al., 2006b). Although it remains unclear, if this is a universal property, adequate removal of the task-induced variance for functional connectivity should yield a correlation profile similar to “continuous” resting state data. A potential method for removing systematic task-induced variance is to model the task effects within a general linear model (GLM) design, remove them, and then analyze the time series of the remaining (residual) signal.

Using correlation analysis, we compared the functional connectivity of resting epochs taken from a mixed/block event-related design fMRI data set and the residuals derived from event-related data with standard continuous resting state data to determine which class of data can best emulate resting state data.

Section snippets

Subjects

All subjects were right-handed native English speakers with normal or corrected-to-normal vision. Subjects were recruited from Washington University and the local community. Participants were screened with a questionnaire to ensure that they had no history of neurological/psychiatric diagnoses or drug abuse. Informed consent was obtained from all subjects in accordance with the guidelines and approval of the Washington University Human Studies Committee. Subjects participated in return for

Removal of task effects from event-related data

Our GLM approach appeared to remove linear task effects. The main effect of time image (ANOVA, Monte Carlo corrected) prior to task removal showed clear task-evoked activity (see Supplementary material). In contrast, the main effect of time image (ANOVA, corrected) for the residual timecourses revealed no significant activations (see Supplementary material).

While the group comparisons between the event-related data residuals and the “continuous” resting state data were qualitatively very

Discussion

In this paper, we examined resting state functional connectivity using two methods that utilize existing event-related and blocked fMRI data. With both methods we attempted to remove the task effects and compare these data to a set of data collected as “continuous” resting state. Both methods produced qualitatively similar results. However, based on the direct comparisons, the interleaved resting state data were superior at emulating “continuous” resting state data.

Acknowledgments

The authors thank the participants in this study, as well as Jessica Church for logistical aid, Mark McAvoy for neuroimaging application development, and David Van Essen and his colleagues for the use of CARET for figures. This work was supported in part by the Washington University Chancellor’s Fellowship and UNCF* Merck Graduate Science Research Dissertation Fellowship to Damien Fair and by NIH NSADA (B.L.S.), NS32979 (S.E.P.), NS41255 (S.E.P.), and NS46424 (S.E.P.), The McDonnell Center for

References (51)

  • M.J. Lowe et al.

    Correlations in low-frequency BOLD fluctuations reflect cortico-cortical connections

    NeuroImage

    (2000)
  • F. Miezin et al.

    Characterizing the hemodynamic response: effects of presentation rate, sampling procedure, and the possibility of ordering brain activity based on relative timing

    NeuroImage

    (2000)
  • F.M. Miezin et al.

    Characterizing the hemodynamic response: effects of presentation rate, sampling procedure, and the possibility of ordering brain activity based on relative timing

    NeuroImage

    (2000)
  • Y. Nir et al.

    Widespread functional connectivity and fMRI fluctuations in human visual cortex in the absence of visual stimulation

    NeuroImage

    (2006)
  • I. Prohovnik et al.

    Observations on the functional significance of regional cerebral blood flow in “resting” normal subjects

    Neuropsychologia

    (1980)
  • L. Tian et al.

    Altered resting-state functional connectivity patterns of anterior cingulate cortex in adolescents with attention deficit hyperactivity disorder

    Neurosci. Lett.

    (2006)
  • B. Biswal et al.

    Functional connectivity in the motor cortex of resting human brain using echo-planar MRI

    Magn. Reson. Med.

    (1995)
  • A.L. Bokde et al.

    Functional connectivity of the fusiform gyrus during a face-matching task in subjects with mild cognitive impairment

    Brain

    (2006)
  • T.T. Brown et al.

    Developmental changes in human cerebral functional organization for word generation

    Cereb. Cortex

    (2005)
  • C. Buchel et al.

    Modulation of connectivity in visual pathways by attention: cortical interactions evaluated with structural equation modelling and fMRI

    Cereb. Cortex

    (1997)
  • C.M. Clark et al.

    Correlational methods for determining regional coupling of cerebral glucose metabolism: a pilot study

    Biol. Psychiatry

    (1984)
  • D.A. Fair et al.

    A comparison of ANOVA and correlation methods for investigating cognitive development with fMRI

    Dev. Neuropsychol.

    (2006)
  • S.D. Forman et al.

    Improved assessment of significant activation in functional magnetic resonance imaging (fMRI): use of a cluster-size threshold

    Magn. Reson. Med.

    (1995)
  • M.D. Fox et al.

    The human brain is intrinsically organized into dynamic, anticorrelated functional networks

    Proc. Natl. Acad. Sci. U. S. A.

    (2005)
  • M.D. Fox et al.

    Spontaneous neuronal activity distinguishes human dorsal and ventral attention systems

    Proc. Natl. Acad. Sci. U. S. A.

    (2006)
  • Cited by (472)

    View all citing articles on Scopus
    View full text