Elsevier

NeuroImage

Volume 102, Part 1, 15 November 2014, Pages 142-151
NeuroImage

The structural–functional connectome and the default mode network of the human brain

https://doi.org/10.1016/j.neuroimage.2013.09.069Get rights and content

Highlights

  • Structure–function connectivity relationship

  • Multi-modal data fusion

  • Voxel-wise connectivity analysis

  • Default mode network

  • Global fiber-tracking

Abstract

An emerging field of human brain imaging deals with the characterization of the connectome, a comprehensive global description of structural and functional connectivity within the human brain. However, the question of how functional and structural connectivity are related has not been fully answered yet. Here, we used different methods to estimate the connectivity between each voxel of the cerebral cortex based on functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI) data in order to obtain observer-independent functional–structural connectomes of the human brain. Probabilistic fiber-tracking and a novel global fiber-tracking technique were used to measure structural connectivity whereas for functional connectivity, full and partial correlations between each voxel pair's fMRI-timecourses were calculated. For every voxel, two vectors consisting of functional and structural connectivity estimates to all other voxels in the cortex were correlated with each other. In this way, voxels structurally and functionally connected to similar regions within the rest of the brain could be identified. Areas forming parts of the ‘default mode network’ (DMN) showed the highest agreement of structure–function connectivity. Bilateral precuneal and inferior parietal regions were found using all applied techniques, whereas the global tracking algorithm additionally revealed bilateral medial prefrontal cortices and early visual areas. There were no significant differences between the results obtained from full and partial correlations. Our data suggests that the DMN is the functional brain network, which uses the most direct structural connections. Thus, the anatomical profile of the brain seems to shape its functional repertoire and the computation of the whole-brain functional–structural connectome appears to be a valuable method to characterize global brain connectivity within and between populations.

Introduction

The analysis of structural and functional connectivity between different regions of the brain provides profound insights into its underlying organization. Thus, both kinds of brain connectivity have been extensively studied over the last years, leading to the concept of the connectome, which was first defined for structural connectivity as a ‘comprehensive structural description of the network of elements and connections forming the human brain’ (Sporns et al., 2005). Later, the concept was extended to include functional connectivity, i.e. the functional connectome (Biswal et al., 2010, Zuo et al., 2011). However, only few studies have focused on the relationship between structural and functional brain connectivity (for reviews see Bassett and Bullmore, 2009, Damoiseaux and Greicius, 2009, Honey et al., 2010, Rykhlevskaia et al., 2008, Sporns, 2011).

Koch et al. (Koch et al., 2002) were first to combine two MRI-imaging techniques to directly study the interplay between structure and function of the human brain. Using diffusion tensor imaging (DTI) and resting-state functional magnetic resonance imaging (fMRI) they provided evidence that fMRI signal correlations reflect the presence of direct and indirect anatomical pathways. Findings from 23 pairs of voxels each situated in two adjacent gyri on a single axial slice were reported. All time courses from structurally connected voxels were also correlated functionally. In some cases, however, functional connectivity was also found in the absence of robust structural connectivity. Building up on this initial point-to-point analysis, a study by (Greicius et al., 2009) found structural connectivity between functionally connected parts of the default mode network (DMN), a set of regions typically active at rest (Buckner et al., 2008, Raichle et al., 2001, Shulman et al., 1997). The precuneus/retrosplenial cortex was found to be structurally connected with bilateral medial temporal lobes (MTL) and the medial prefrontal cortex (MPFC). Tracts starting from the medial prefrontal cortex (MPFC) contacted the posterior cingulate cortex (PCC).

A similar study extended these qualitative findings to a quantitative level (van den Heuvel et al., 2008). Partial correlations between the PCC and the MPFC (regressing out 15 other clusters within the DMN and other resting-state networks) accounted for functional connectivity strength, whereas mean fractional anisotropy values of the cingulum were used as a measure for structural connectivity strength. The two measures of connectivity strength correlated between subjects, which suggests that a strong structural basis leads to strong functional couplings between regions.

A limitation of region-to-region connectivity measurements, as applied in the studies above, is that regions of interest must be manually selected, possibly introducing selection bias (Damoiseaux and Greicius, 2009). To avoid this, data-driven approaches such as independent component analysis (ICA) have become increasingly common, allowing for the analysis of functional connectivity across the whole brain. A study by (Segall et al., 2012) and colleagues simultaneously applied ICA on functional time courses and gray matter density values (Xu et al., 2009) to analyze the global structure–function relationship. High spatial overlap between functional and structural ICA components was found. However, in the ICA framework, the number of independent components has to be defined typically a priori and the technique cannot be applied easily to anatomical white-matter connectivity.

A different approach that allows for analyzing whole-brain connectivity without a manual definition of regions of interest is to measure connectivity between each pair of voxels within the gray matter. To the best of our knowledge, only few studies have used such an approach and only two have focused on the relationship between structure and function in a voxel-wise (Skudlarski et al., 2008) or more coarsely parcellated fashion (Honey et al., 2009).

In the study by Honey et al., the brain was initially divided into 66 regions based on landmarks using FreeSurfer (Fischl et al., 2004). These regions were then further subdivided into 998 areas of approximately equal size. Functional connectivity was measured by correlating the mean fMRI signals of each area. The number of fiber tracts connecting regions was used as a measure or their structural connectivity strength. It was found that both indices of connectivity strength correlated between regions. However, as shown in the studies by Koch et al. and Greicius et al., functional connectivity was also found between regions that were not directly connected by anatomical tracts. In the study by Skudlarski and colleagues, the agreement between structure and function was analyzed in a voxel-wise fashion. Thereby, functional and structural connectivity between each of 5000 downsampled isotropic 4 mm voxels within the gray matter was compared. Strongest correlations between functional and structural connectivity were found in highly connected regions such as the thalamus, precuneus and anterior cingulate.

Based on the work by Skudlarski et al., in the current study we estimated structural and functional connectivity at high spatial resolution between each pair of 40,000 voxels, defined by a group-template that best represented the mean anatomy of all participants of the study. For each voxel within the cortex, structural and functional connectivity to all other voxels was estimated. Thus, connectivity vectors that accounted for different connectivity estimates of a certain voxel to the rest of the brain were established. In a second step, these were correlated to each other to identify voxels that were both structurally and functionally connected to similar regions within the rest of the brain. In this way, a parcellation-independent processing stream was established that makes it possible to study the relationship of structure and function within and between populations.

Furthermore, to study the impact of different methods on the results and the robustness of the approach, we compared two functional and two structural connectivity measures. Full and partial correlations were estimated to assess functional connectivity, whereas probabilistic and global fiber-tracking was performed to analyze structural connectivity. In a comparison study of different network modeling techniques (Smith et al., 2011) and in direct comparison with structural equation modeling (Marrelec et al., 2009), partial correlations performed well in reconstructing networks based on time series data. Partial correlations are commonly used to analyze direct relationships between time signals, whereas full correlations are more prone to introduce indirect connections into the model. To analyze the difference between a method focusing on direct connectivity and one that largely introduces indirect connectivity, we therefore compared estimates of partial and full correlations of functional time series in this study. To estimate structural connectivity, we used a classical, probabilistic fiber-tracking method based on a diffusion tensor model (Kreher et al., 2008) and a novel, global fiber-reconstruction technique that has been shown to perform well in the reconstruction of crossing fibers (Reisert et al., 2011). The results for the different methods were strikingly similar, revealing most parts of the default mode network, a functional network that consists of the precuneus and adjacent posterior cingulate/retrosplenial cortex (PCC/Rsp), the medial prefrontal cortex (MPFC) and inferior parietal lobes/angular gyrus (IPL/AG) as well as the medial temporal lobe (MTC). Thus, a voxel-wise comparison of structural and functional connectivity appears useful to characterize the functional and anatomical architecture of the human brain and seems to yield robust results even when using different methods.

Section snippets

Participants

Nineteen healthy subjects (6 females, 18 right-handed) between 21 and 31 years of age (mean: 26.6, standard deviation: 2.98) participated in the study. All measurements were obtained on a 3 Tesla whole-body scanner (Trio, Siemens, Erlangen, Germany) using a birdcage radio-frequency head coil. Subjects had no neurological or psychiatric history and were not taking any psychoactive medication. The study was approved by the local ethics committee of the Charité, University Medicine Berlin and

Overall matrix agreement

Mean matrix agreement between functional and structural whole-brain connectivity matrices was assessed by calculating the Pearson's correlation coefficient between two connectivity matrices. Following the approach of (Skudlarski et al., 2008), this was done after removing the effect of Euclidean distance, i.e. by calculating the partial correlation coefficients between structural and functional connectivity matrices not accounted for by the effect of Euclidean distance. However, to be able to

Discussion

The primary aim of this study was to introduce a method that makes it possible to investigate the structural and functional connectivity of the whole brain simultaneously. To achieve this, an observer-independent processing stream was established that computes voxel-to-voxel connectivity within the entire cerebral cortex. Two methods to estimate functional, as well as structural connectivity were compared. For all combinations of methods, areas of the default mode network showed the highest

Acknowledgments

This work was supported by a grant from the German Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung; F.B.) and by the German Research Foundation (DFG), KFO247. We would like to thank Mark Schmidt, Ryszard Auksztulewicz and Timo Schmidt for their helpful advice.

Conflict of interest statement

We hereby certify that there is no conflict of interest regarding the material discussed in the manuscript.

References (64)

  • P. Skudlarski et al.

    Brain connectivity is not only lower but different in schizophrenia: a combined anatomical and functional approach

    Biol. Psychiatry

    (2010)
  • S.M. Smith et al.

    Network modelling methods for FMRI

    NeuroImage

    (2011)
  • A. Venkataraman et al.

    Whole brain resting state functional connectivity abnormalities in schizophrenia

    Schizophr. Res.

    (2012)
  • A. Weissenbacher et al.

    Correlations and anticorrelations in resting-state functional connectivity MRI: a quantitative comparison of preprocessing strategies

    NeuroImage

    (2009)
  • L. Xu et al.

    Joint source based morphometry identifies linked gray and white matter group differences

    NeuroImage

    (2009)
  • S. Achard et al.

    A resilient, low-frequency, small-world human brain functional network with highly connected association cortical hubs

    J. Neurosci.

    (2006)
  • J. Alstott et al.

    Modeling the impact of lesions in the human brain

    PLoS Comput. Biol.

    (2009)
  • S.P. Awate et al.

    Fuzzy nonparametric DTI segmentation for robust cingulum-tract extraction

    Med Image Comput. Comput. Assist. Interv.

    (2007)
  • O. Banerjee et al.

    Convex optimization techniques for fitting sparse Gaussian graphical models

  • D.S. Bassett et al.

    Human brain networks in health and disease

    Curr. Opin. Neurol.

    (2009)
  • D.S. Bassett et al.

    Hierarchical organization of human cortical networks in health and schizophrenia

    J. Neurosci.

    (2008)
  • B.B. Biswal et al.

    Toward discovery science of human brain function

    Proc. Natl. Acad. Sci.

    (2010)
  • R.L. Buckner et al.

    The brain's default network: anatomy, function, and relevance to disease

    Ann. N. Y. Acad. Sci.

    (2008)
  • R.L. Buckner et al.

    Cortical hubs revealed by intrinsic functional connectivity: mapping, assessment of stability, and relation to Alzheimer's disease

    J. Neurosci.

    (2009)
  • V.D. Calhoun et al.

    Exploring the psychosis functional connectome: aberrant intrinsic networks in schizophrenia and bipolar disorder

    Front. Psychiatry

    (2011)
  • J.S. Damoiseaux et al.

    Greater than the sum of its parts: a review of studies combining structural connectivity and resting-state functional connectivity

    Brain Struct. Funct.

    (2009)
  • G. Deco et al.

    Emerging concepts for the dynamical organization of resting-state activity in the brain

    Nat. Rev. Neurosci.

    (2011)
  • B. Fischl et al.

    Automatically parcellating the human cerebral cortex

    Cereb. Cortex

    (2004)
  • P.T. Fletcher et al.

    A volumetric approach to quantifying region-to-region white matter connectivity in diffusion tensor MRI

    Inf. Process. Med. Imaging

    (2007)
  • K.J. Friston et al.

    Statistical parametric maps in functional imaging: a general linear approach

    Hum. Brain Mapp.

    (2004)
  • A.G. Garrity et al.

    Aberrant “default mode” functional connectivity in schizophrenia

    Am. J. Psychiatry

    (2007)
  • G. Gong et al.

    Mapping anatomical connectivity patterns of human cerebral cortex using in vivo diffusion tensor imaging tractography

    Cereb. Cortex

    (2009)
  • Cited by (247)

    View all citing articles on Scopus
    View full text