0
We're unable to sign you in at this time. Please try again in a few minutes.
Retry
We were able to sign you in, but your subscription(s) could not be found. Please try again in a few minutes.
Retry
There may be a problem with your account. Please contact the AMA Service Center to resolve this issue.
Contact the AMA Service Center:
Telephone: 1 (800) 262-2350 or 1 (312) 670-7827  *   Email: subscriptions@jamanetwork.com
Error Message ......
Original Investigation |

Abnormal Rich Club Organization and Functional Brain Dynamics in Schizophrenia FREE

Martijn P. van den Heuvel, PhD1; Olaf Sporns, PhD2; Guusje Collin, MD1; Thomas Scheewe, PhD1; René C. W. Mandl, PhD1; Wiepke Cahn, MD, PhD1; Joaquín Goñi, PhD2; Hilleke E. Hulshoff Pol, PhD1; René S. Kahn, MD, PhD1
[+] Author Affiliations
1Department of Psychiatry, Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, Utrecht, the Netherlands
2Department of Psychological and Brain Sciences, Indiana University, Bloomington
JAMA Psychiatry. 2013;70(8):783-792. doi:10.1001/jamapsychiatry.2013.1328.
Text Size: A A A
Published online

Importance  The human brain forms a large-scale structural network of regions and interregional pathways. Recent studies have reported the existence of a selective set of highly central and interconnected hub regions that may play a crucial role in the brain’s integrative processes, together forming a central backbone for global brain communication. Abnormal brain connectivity may have a key role in the pathophysiology of schizophrenia.

Objective  To examine the structure of the rich club in schizophrenia and its role in global functional brain dynamics.

Design  Structural diffusion tensor imaging and resting-state functional magnetic resonance imaging were performed in patients with schizophrenia and matched healthy controls.

Setting  Department of Psychiatry, Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, Utrecht, the Netherlands.

Participants  Forty-eight patients and 45 healthy controls participated in the study. An independent replication data set of 41 patients and 51 healthy controls was included to replicate and validate significant findings.

Main Outcome(s) and Measures  Measures of rich club organization, connectivity density of rich club connections and connections linking peripheral regions to brain hubs, measures of global brain network efficiency, and measures of coupling between brain structure and functional dynamics.

Results  Rich club organization between high-degree hub nodes was significantly affected in patients, together with a reduced density of rich club connections predominantly comprising the white matter pathways that link the midline frontal, parietal, and insular hub regions. This reduction in rich club density was found to be associated with lower levels of global communication capacity, a relationship that was absent for other white matter pathways. In addition, patients had an increase in the strength of structural connectivity–functional connectivity coupling.

Conclusions  Our findings provide novel biological evidence that schizophrenia is characterized by a selective disruption of brain connectivity among central hub regions of the brain, potentially leading to reduced communication capacity and altered functional brain dynamics.

Figures in this Article

The human brain is a complex network of structurally and functionally interconnected regions. Studies14 examining the brain’s underlying network structure are motivated by the notion that brain function is not solely attributable to the properties of individual regions or individual connections but rather emerges from the network organization of the brain as a whole, the human connectome.3,58 Conversely, brain dysfunction may result from abnormal wiring of the brain’s network.9,10

The notion that schizophrenia, a severe psychiatric disorder characterized by hallucinations, delusions, loss of initiative, and cognitive dysfunction, may relate to disconnectivity among brain regions has a long history. As cited by Stephan,11 Wernicke was among the first to suggest that schizophrenia may involve anatomical disruption of association pathways. Bleuler,12 who coined the term schizophrenia, hypothesized that decoupling of psychological processes might be the primary cause of the disease. Lately, studies using imaging techniques, such as diffusion tensor imaging (DTI) and functional magnetic resonance imaging (fMRI), have reported widespread disconnectivity in patients, in particular reduced integrity of frontal and temporal white matter connections1319 and affected functional coupling of the default mode network.2025

A few highly connected and central regions, the so-called hub nodes, have a key role in the global topology of the brain’s network.1,2630 Previous network studies29,3135 report disruptions in the overall organization of structural connectivity (SC) and functional connectivity (FC) in patients with schizophrenia, together with a less centralized position of some of these hubs in the frontal, temporal, and parietal cortex.9 However, it remains unknown whether reduced connectivity of hubs constitutes a nonspecific generalized phenomenon involving white matter connectivity to and from all brain regions or whether this disruption disproportionally involves pathways that link highly connected regions. In the healthy brain, hubs have been found to be densely interconnected, together forming a central core or rich club,1,36,37 with rich club connections having a pivotal role in interregional brain communication.38 We test the hypothesis that disturbed wiring of this central rich club may contribute to the pathophysiology of schizophrenia.

Using neuroimaging data in a group of 48 patients and 45 healthy controls, we examined potentially abnormal connectivity of the brain’s rich club and the relationship of a disruption of this communication backbone to (reduced) levels of global communication capacity and changed functional dynamics in the brain networks of patients. An independently acquired data set of patients and controls (41 patients and 51 controls) was used to replicate possible findings.

Two data sets were included: a principal data set of 48 patients and 45 healthy controls collected with 3-T MRI and an independent replication data set of 41 patients and 51 matched controls acquired with 1.5-T MRI. Demographics of the data sets are listed in the Table (see also the eMethods in Supplement for a detailed description of the demographics). All participants provided written informed consent. Participants underwent psychiatric assessment procedures at the Department of Psychiatry, Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, Utrecht, the Netherlands, using the Comprehensive Assessment of Symptoms and History. Diagnostic consensus of patients was achieved in the presence of a psychiatrist with the DSM-IV criteria for schizophrenia.

Table Graphic Jump LocationTable.  Demographic and Clinical Characteristics of Principal and Replication Data Sets
Data Acquisition and Analysis
Principal Data Set

The principal data set was acquired with a 3-T scanner (Philips Achieva). T1, DTI (2 sets of 32 weighted directions), and resting-state fMRI data (8 minutes) were acquired (eMethods in Supplement).3941

Replication Data Set

The replication data set was acquired with a 1.5-T clinical scanner (Philips Achieva) and included a T1, DTI (2 sets of 30 directions), and resting-state fMRI scan (8 minutes).29,41,42

Preprocessing

Data processing steps are illustrated in Figure 1 (see eMethods in Supplement for a detailed description). First, the gray matter cortex was parcellated into distinct regions using the Freesurfer software suite (Athinoula A. Martinos Center for Biomedical Imaging at Massachusetts General Hospital) (Figure 1A), including 68 cortical regions and 14 subcortical regions.43 Second (Figure 1B), DTI scans were realigned and corrected for motion and susceptibility distortions. White matter pathways, referred to as fibers or tracts, were reconstructed using streamline tractography. Third (Figure 1C), resting-state time series were realigned, coregistered with the T1 image, detrended (linear trends and first-order drifts were removed), corrected for global effects (regressing out 6 motion parameters, white matter, ventricle, and global mean signals), and band-pass filtered (0.01 to 0.1 Hz). Motion correction on the time series44,45 and analyzing the functional data without global mean correction46,47 did not change the nature of the results reported (see eResults in Supplement).

Place holder to copy figure label and caption
Figure 1.
The Analysis Steps

A, Individual T1 images were used for automatic classification of gray and white matter tissue and parcellation of the cortex into 68 distinct brain regions, forming the nodes of the individual brain networks. B, Streamline tractography was applied to the diffusion tension imaging (DTI) data to reconstruct cortico-cortical white matter pathways. From the set of reconstructed streamlines, streamlines that interconnected region i and j from the set of 68 or 82 regions were taken as an edge between node i and j in the structural brain network. Streamline count was taken to represent the weight of the connection and was aggregated into a structural connectivity (SC) matrix. C, Functional connectivity (FC) between node i and j was computed as their level of correlation between their resting-state function magnetic resonance imaging (fMRI) blood oxygenation level dependent (BOLD) time series, resulting in a matrix FC. D, The topologic organization of the resulting individual structural brain networks was examined, including (among other metrics) measurements of the rich club, global strength, and global efficiency (top). The level of coupling between SC and FC was examined by computing the level of correlation between the weights of (existing) structural connections and their functional counterparts. This correlation is referred to as the level of SC-FC coupling. At the group level, values were examined between the patient and control groups (for both the principal and replication data sets). Statistical evaluation was performed using permutation testing (see Methods and eMethods in Supplement).

Graphic Jump Location
Connectivity Analysis

For each study participant an SC and an FC network were created as follows.

Construction of Structural Brain Networks

For each data set a structural brain network was created by combining the collection of reconstructed fiber tracts with the individual parcellation map (Figure 1B, eMethods in Supplement). A network consists of a set of nodes and connections that can be mathematically expressed as a graph: G = (V,E), with V indicating the collection of nodes and E the collection of edges between the nodes. Two networks were created: (1) a cortical network consisting of 68 cortical regions and (2) a whole-brain network consisting of the 68 cortical and 14 subcortical regions. Network nodes i and j were defined as being structurally connected when a set of fibers was found from the total collection of reconstructed streamlines that interconnected them (Figure 1B). In addition to analyzing weighted networks based on absolute streamline count, an analysis was performed in which the weights of the connections were expressed as streamline densities, computed as the number of streamlines divided by the individual volumes of the 2 interconnected regions of interest.1,36

Construction of Functional Brain Networks

For each data set, the level of FC between each pair of nodes in the network was computed as the correlation between their averaged regional time series (Figure 1C, eMethods in Supplement). The nodes of the functional network were taken in a similar manner as the nodes in the structural network to enable direct comparison between the 2 types of networks.

Graph Analysis of Structural Connectome Topology
Graph Organizational Metrics

Characteristic graph metrics were computed to examine possible differences in overall connectome topology between patients and controls. Graph metrics are described in detail elsewhere.36,48 Examined metrics included the degree strength (computed as the sum of the weights of the node’s connections), the clustering coefficient (reflecting the level of local connectedness of a node), the shortest path length (reflecting the mean minimal travel distance between nodes in the network), the global efficiency (reflecting the capacity for network-wide communication), and the modularity (reflecting the level of community structure in the network49). All metrics were computed based on the individually weighted (streamline count) networks. Clustering and length were compared with the clustering coefficient and shortest path length of a set of random networks (n = 1000),48 providing a normalized clustering coefficient γ and normalized path length λ, respectively.

Rich Club Organization of the Structural Brain Network

The so-called rich club phenomenon in networks is said to be present when the highly connected (high-degree) hubs of a network are more densely connected among themselves than predicted on the basis of their high degree alone.50 Rich club organization of the human connectome was found in a previous study.36 A weighted version of the rich club behavior incorporates the weights of the edges in the network,51 examining the level of density between the subset of selected nodes in the network. A formal description of the weighted rich club coefficient is given in the eMethods in Supplement. In short, the weighted rich club coefficient Φw(k) is computed as the sum of the weights of the subset of connections E>k of the nodes with a degree >k in the network divided by the sum of the set of the strongest E>k connections in the total network. Φw(k) is typically examined relative to the averaged rich club curve Φw random(k) of a (set of) comparable random network(s) to determine the extent to which empirically observed connection density between rich club nodes exceeds that predicted by a random null model.50,52 A normalized coefficient Φwnorm(k) (given as the ratio Φw/Φrandomw) of greater than 1 over a range of k suggests the existence of rich club organization in a network.50,51 For each network, a population of m = 1000 random networks48 was computed by shuffling the links in M, preserving the weights and the degree sequence and thus all node degrees (including the hubs) in the network. For convenience, for the rest of this article Φwnorm(k) will be referred to as Φw(k).

Rich Club Nodes

Rich club regions as described throughout this article were selected on the basis of the group-averaged cortical network, set at a rich club level of k>15, as previously reported.36 Selecting the rich club regions on the group level was consistent with selecting rich club regions on the individual level by ranking for each data set the nodes on the basis of rich club level k and selecting the top 12% of most consistently ranked nodes across the group of subjects (Results).36 Rich club effects as reported in this article were also present at other group levels of k, as well as when the rich club was selected on the individual level (eResults in Supplement).

Connection Classes: Rich Club, Feeder, and Local Connections

On the basis of the categorization of the nodes of the network into rich club and non–rich club regions, edges of the network were classified into rich club connections, linking rich club nodes to rich club nodes; feeder connections, linking, rich club nodes to non–rich club nodes; and local connections linking non–rich club nodes to non–rich club nodes (Figure 2C).38

Place holder to copy figure label and caption
Figure 2.
Rich Club Organization

A, Group-averaged rich club curve for controls (white) and patients (blue). Patients had a significantly reduced rich club organization for the range k = 26 to k = 28, reflecting a lower level of connectivity between central hubs of the brain (cortical: P = .003; whole brain: P = .04; 10 000 permutations) B, Confirming previous findings, rich club members included the bilateral precuneus, superior frontal cortex, superior parietal cortex, and the insula in both the healthy and patient populations. This figure is based on the group-averaged cortical network in controls (at a rich club level of k >15).53 C, Edges across individual brain networks (both controls and patients) were divided into 3 distinct classes: rich club connections linking rich club members (red), feeder connections linking rich club members to non–rich club members (orange), and local connections connecting non–rich club members (yellow edges). Examining the density of rich club, feeder, and local connections between the populations of controls and patients revealed a significant reduction in rich club density in patients but no significant effect in density of feeder and local connections. The figure shows mean (SD) density values for each of the 3 classes, scaled to the mean density values of the control group. ROI indicates region of interest.

Graphic Jump Location
Analysis of Functional Coupling

Individual functional networks were examined for their node-specific level of connectivity strength (ie, the total sum of connectivity strength of a node) and their level of modularity.48

SC-FC Relationship

For each network, a correlation analysis was performed between the strength of the structural connections and their functional counterparts. All nonzero entries of the SC matrix were selected, rescaled to a gaussian distribution, and correlated with their functional counterparts selected from the FC matrix. This resulted in a single SC-FC coupling metric for each of the brain networks.54 Rescaling of the structural weights to a gaussian distribution1,54 was used to normalize the distribution of SC values. Rescaling did not change the nature of our findings (eResults in Supplement).

Statistical Analysis

To evaluate the statistical relevance of observed effects (eg, rich club coefficient, rich club, feeder, local connectivity density, and graph metrics), permutation testing was used to randomize group assignment (eMethods in Supplement for a detailed description).29,55 Permutation testing yielded an empirical null distribution of effects under the null hypothesis that patient and control groups were not different, assigning the difference between the patient and control group a (2-tailed) P value by determining the percentage of the computed null distribution that exceeded the empirically measured metric.

Rich Club Organization
Principal Data Set

Rich club organization was found in the structural networks of both the healthy controls and the patient group. Rich club members included bilateral precuneus, superior frontal cortex, superior parietal cortex, and the insula (Figure 2B).

Rich Club

Figure 2A illustrates the group averaged normalized Φw rich club curves of both the control and patient networks (based on streamline count). Patients had a significantly reduced rich club organization, reflecting a lower level of connectivity between central hubs of the brain (cortex: P = .03 for rich club range k = 16 to k = 28); a marginal effect was found for whole brain networks: P = .04 range k = 28 to k = 29; 10 000 permutations). Dividing off the regional volume (correcting for possible regional volumetric effects) did not change the nature of our findings (cortex: P = .01; whole brain: P = .10; 10 000 permutations). Notably, reduced rich club organization in the patients was most pronounced in the cortical networks, suggesting that connectivity is perturbed between cortical hubs in schizophrenia.

Replication Data Set

Confirming abnormal rich club organization in patients, the replication data set demonstrated a reduced Φw regimen in the cortical networks (cortex: for k = 15 to k = 20, P = .003; whole brain: for k = 27, P = .18; both 10 000 permutations, Figure 3A).

Place holder to copy figure label and caption
Figure 3.
Replication Data Set

A, Group-averaged rich club curve of controls (white) and patients (blue) of the replication data set (n = 41 patients and 51 controls). B, Global graph metrics (cortex network) of controls (dark blue) and patients (light blue). No differences were found between patients and controls on connectivity strength (S), global efficiency (GE), or clustering (C). mod indicates modularity. C, Density of rich club, feeder, and local connections. Confirming the findings of the principal data set, patients had a significant reduction in rich club density (*P < .05). D, Structural connectivity (SC)–functional connectivity (FC) coupling for patients and controls (data of 39 patients and 35 controls). Confirming the findings of the principal data set, patients had an increased level of SC-FC coupling.

Graphic Jump Location
Density of Rich Club, Feeder, and Local Connections
Principal Data Set

Further examining reduced rich club interconnectedness, statistical testing revealed a reduced level of SC density (ie, streamline count dividing off regional volume) of rich club connections (ie, connections that link the group-averaged rich club members) in patients (P = .01, permutation testing, 10 000 permutations, Figure 2D). In contrast, this effect was nonexistent for feeder connections and local connections (P = .27 and P = .99, respectively) (Figure 2D). Selection of rich club nodes on the individual level, instead of a priori selection of the rich club nodes on the basis of the group data, revealed similar results (eResults in Supplement). To provide insight into whether abnormal connectivity might be concentrated to the rich club, the ratios of densities between rich club and feeder connections, as well as the ratio of densities between rich club and local connections, were tested between patients and controls (eResults in Supplement). Statistical testing revealed that both ratios were significantly lower in patients compared with controls (P = .04 for rich club–feeder connections and P = .02 for rich club–local connections). Rich club differences were also present when patients with the diagnosis of schizoaffective disorder were excluded (eResults in Supplement).

Replication Data Set

Structural density of rich club connections was significantly reduced in patients (P = .04, 19% reduction), whereas no difference was found in structural density of feeder connections (P = .17). In the replication data set, density of local connections was found to be statistically different from controls but less pronounced than the reduction in rich club density (P = .04, 14% reduction, Figure 3C).

Other Graph Metrics: Structural Networks
Principal Data Set

Connectivity strength was found to be reduced in patients (weights based on streamline count; cortex: P = .04; whole brain P = .03, 10 000 permutations) (Figure 4A). Both controls and patients had a small-world organization of the structural connectome, showing a high level of local clustering (normalized clustering-coefficient γ>>1) and a high level of global integration (normalized path length λ of  approximately  1). Global efficiency was found to be reduced in patients (cortex: P = .02; whole brain: P = .02). Overall clustering was found to be slightly reduced in patients (cortex: P = .04; whole brain P = .02). Brain networks of patients did not differ from controls for the normalized clustering coefficient γ or normalized path length λ (eMethods in Supplement), supporting the notion of a preserved overall small-world topology, as previously reported.29,56 Brain networks of patients revealed an elevated level of structural modularity compared with the healthy controls, which reached statistical significance for whole-brain networks (cortex: P = .10; whole brain: P = .03). Dividing off effects of regional volume and correcting for global strength revealed a marginal decrease in overall global efficiency of patients compared with controls (cortex: P = .07; whole brain: P = .002, 10 000 permutations). No significant differences in strength or clustering were found when density-weighted networks (ie, number of streamlines dividing off regional volume) were examined.

Place holder to copy figure label and caption
Figure 4.
Global Efficiency and Rich Club Density

A, General topologic graph metrics. Patients had reduced levels of connectivity strength (S) and global efficiency (GE) and increased levels of local clustering (C) and modularity (mod) (*P < .05, permutation testing, 10 000 permutations, effects of GE remained significant after volume and global S correction, cortex network). B, Across the control (top) and patient (bottom) populations, global efficiency (y-axis) was significantly correlated with rich club density (x-axis), after correcting for overall differences in global connectivity S.

Graphic Jump Location
Replication Data Set

Consistent with the principal data set, global efficiency (cortex: P = .02; whole brain: P = .02) and overall connectivity strength were reduced in patients (cortex: P = .03; whole brain: P = .03). Clustering was found to be reduced in patients (cortex: P = .004; whole brain P = .01; Figure 3B).

Rich Club Density and Global Efficiency
Principal Data Set

Global efficiency significantly correlated with overall connectivity strength and rich club connectivity in both the control and patient populations. Correcting for individual variation in overall connectivity strength (regressing out over the total group of controls and patients) revealed a significant association between rich club density (ie, streamline count dividing off regional volume) and global efficiency in both the patient (R = 0.45, P = .001) and control populations (R = 0.328, P = .03; Figure 4B), as well as over the 2 groups combined (R = 0.42, P < .001). The patient group revealed a marginal higher correlation, but this group interaction effect was not statistically significant (P = .40).

Replication Data Set

Global efficiency and rich club connectivity (corrected for overall connectivity strength) revealed a positive association at trend level for patients (R = 0.26; P = .06) and controls (R = 0.29; P = .06)

SC-FC Coupling
Principal Data Set

Patients had a significantly reduced level of overall FC strength (P < .001, 10 000 permutations), supporting previous fMRI results.32 Healthy controls had a moderate correlation between SC and FC.1,32 The SC-FC coupling for the cortical networks was found to be 0.25 (SD, 0.048) in the control group, whereas SC-FC coupling in patients was found to be significantly increased to 0.28 (SD, 0.044) (Figure 5A), suggesting an increase in SD-FC coupling in patients relative to controls (cortex: P = .04; mean [SD] for whole brain: controls, 0.19 [0.049]; patients, 0.21 [0.044]; P = .02; 10 000 permutations). Increased SC-FC coupling was found to be present in local connections, with feeder and rich club connections showing no significant difference in SC-FC coupling (eResults in Supplement).

Place holder to copy figure label and caption
Figure 5.
Structural Connectivity (SC)–Functional Connectivity (FC) Coupling

A, Patients revealed a significant increase in SC-FC coupling compared with healthy controls (*P < .05, permutation testing, 10 000 permutations). B, In the patient population (bottom), lower levels of normalized rich club density (measured as the fraction of streamlines inside the rich club relative to the total number of streamlines, corrected for volume effects, blue) were associated with increased SC-FC coupling. This association was absent in the control population (top).

Graphic Jump Location
Replication Data Set

The replication data set confirmed an increased level of SC-FC coupling in patients in cortical networks (cortex: controls, 0.25 [0.062]; patients, 0.28 [0.065]; P = .04; whole brain: controls, 0.18 [0.051]; patients, 0.21 [0.060]; P = .07; 10 000 permutations; Figure 3D).

Association Between Rich Club Density and SC-FC Coupling in Patients
Principal Data Set

A moderate negative correlation was found between rich club density (as a percentage of overall connectivity) and SC-FC in the patient population (cortex: R = −0.27, P = .04, linear regression, 1-sided) (Figure 5B). This correlation was absent in the control group (R = 0.026, P = .87). However, correlations between patients and controls were not statistically different (P = .11).

Replication Data Set

A moderate negative correlation at trend level was found in the patient population (cortex: R = −0.26; P = .05, linear regression, 1-sided), partially confirming the results of the principal data set. This association was absent in the control group (R = 0.18, P = .31). Correlations were not statistically different (P = .19).

Clinical Metrics

No clear association between disease symptoms (eg, Positive and Negative Syndrome Scale and Comprehensive Assessment of Symptoms and History scores) and metrics of rich club organization and/or other graph metrics were found.

The main finding of this study is a reduced level of rich club interconnectivity in patients with schizophrenia. Connectivity density of rich club connections was found to be significantly affected in both the principal and replication data sets, indicative of schizophrenia to be associated with a disturbance in SC among key hubs of the human brain. The rich club takes a central position in the brain’s network topology,36,57 and connections among rich club brain hubs have been proposed to be central to the integration of information among different subsystems of the human brain.38 Our findings suggest that schizophrenia is characterized by reduced structural integrity of this centrally embedded rich club backbone, potentially resulting in decreased global communication capacity and altered functional brain dynamics.

Our findings of reduced rich club connectivity agree with emerging evidence of connectome abnormalities in patients, suggesting that disrupted global brain communication has a key role in the pathophysiology of schizophrenia.9,10,35 First, findings of FC studies suggest that schizophrenia may be related to an aberrant hub role of frontal brain regions,31,32 providing empirical evidence for the idea that schizophrenia may be related to inefficient integration of information among brain regions. Second, our rich club findings (Figure 2) tend to show clear overlap with a reported subnetwork of affected anatomical pathways in schizophrenia,56 consistent with reports of less central and less structurally connected frontal and parietal hubs in patients.29,55 Indeed, DTI studies have suggested a role of frontal-parietal tracts (eg, cingulum) in the disease pathology of schizophrenia,1618,58,59 including anatomical pathways that are suggested to have a role in the rich club.60 Our findings may thus further extend the notion of affected hub connectivity in schizophrenia, raising the hypothesis that connections among brain hubs might be disproportionally affected. Indeed, significant reductions in SC were predominantly found for rich club connections, whereas other structural pathways (ie, feeder and local connections) seem to be relative spared in patients. Our findings may thus suggest that disconnectivity in schizophrenia may, to some extent, be concentrated to connections that link the constituents of the brain’s rich club rather than equally affecting all white matter connections of frontal and temporal brain regions. However, in this context other temporal and frontal (ie, non–rich club) pathways have also been suggested to have a role in the pathophysiology of schizophrenia.16,41,61 Future studies are needed to examine whether and if so to what extent disconnectivity effects in schizophrenia are disproportionally accounted by rich club connections.

Another main observation is an increased coupling between SC and FC in patients with schizophrenia. Although the overall level of both SC and FC was significantly reduced in patients, supporting previous observations and suggesting a reduced overall level of functional interactions in patients,31,32 the coupling between SC and FC was higher in patients. This increased correlation may suggest that the illness leads to functional interactions that are more directly related to the underlying anatomical connectivity of the brain and may be indicative of more stringent and less dynamic brain function in patients. In addition, confirming previous findings,29,31,56 brain networks of patients had reduced levels of structural global efficiency. Our results may add to these findings, suggesting that this reduction in overall global communication capacity might be related to a decrease in rich club density. Rich club connections have been hypothesized to form a backbone for global brain communication,36,38,62 and perhaps disease-induced reductions in rich club density may relate to a reduced structural capacity to integrate information (ie, less global efficiency) among different regions of the brain, resulting in more functionally isolated subsystems. Our findings may thus converge on the notion that schizophrenia is related to reduced structural interconnectivity among rich club hubs accompanied by reduced levels of global communication capacity and affected functional dynamics. Because of the cross-sectional design of our study, the direct causal relationship between abnormal rich club organization and concomitant changes in global efficiency and brain dynamics remains, however, unclear. Graph modeling studies could provide more insight into a possible causal relationship between affected rich club organization and global brain efficiency.63

This study did not reveal a clear association between clinical metrics of patients and rich club organization, suggesting a complex relationship between connectome abnormalities in schizophrenia and clinical symptoms. However, in the context of an absence of such an association, one could speculate about the notion that connectome abnormalities (and rich club abnormalities in particular) might be more related to aspects of global outcome of patients rather than symptom severity or, alternatively, reflect a potential vulnerability factor for the disease. Future studies examining a possible familial and genetic background of rich club abnormalities are of potential importance.10,35

A limitation of the present study involves the use of the number of tractography streamlines among brain regions as the main measure of SC, including the assessment of disease-related differences. A reduction in fiber streamlines may relate to multiple factors, including changes in complex fiber architecture and changes in the local microstructural organization of white matter tissue. In addition, the brain network was examined at a relative low spatial resolution, including cortical regions of varying sizes, an effect that could potentially (eResults in Supplement) confound the observed rich club organization. Other studies30,34,36 have advocated for the use of higher-resolution parcellation schemes, but the optimal resolution for patient studies remains unknown. Small connectivity changes may be missed at lower resolutions because of spatial averaging, but more fine-grained parcellation schemes have been found to involve higher levels of intersubject variability, levels that may outrange differences among groups.64 New advances in multiscale approaches, examining connectome differences across different resolutions, are therefore of interest.64,65

In summary, this study finds abnormal rich club connectivity in schizophrenia, suggesting a reduced level of SC among key hubs of the human brain. Rich club connections have been proposed to constitute a backbone for brain communication, and our findings suggest that schizophrenia may involve a reduced structural integrity of this centrally embedded communication backbone. Reduced integrity of this system in schizophrenia may contribute to altered brain dynamics and reduced integration of information among different systems of the human brain.

Corresponding Author: Martijn Pieter van den Heuvel, PhD, Department of Psychiatry, University Medical Center Utrecht, Rudolf Magnus Institute of Neuroscience, Heidelberglaan 100, PO Box 85500, 3508 GA Utrecht, The Netherlands (m.p.vandenheuvel@umcutrecht.nl)

Submitted for Publication: June 28, 2012; final revision received September 29, 2012; accepted December 1, 2012.

Published Online: June 5, 2013. doi:10.1001/jamapsychiatry.2013.1328.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by VENI grant 451-12-001 from the Netherlands Organization for Scientific Research (Dr van den Heuvel) and by the J.S. McDonnell Foundation (Drs Sporns and Goñi).

Hagmann  P, Cammoun  L, Gigandet  X,  et al.  Mapping the structural core of human cerebral cortex. PLoS Biol. 2008;6(7):e159.
PubMed   |  Link to Article
Sporns  O, Chialvo  DR, Kaiser  M, Hilgetag  CC.  Organization, development and function of complex brain networks. Trends Cogn Sci. 2004;8(9):418-425.
PubMed   |  Link to Article
van den Heuvel  MP, Stam  CJ, Kahn  RS, Hulshoff Pol  HE.  Efficiency of functional brain networks and intellectual performance. J Neurosci. 2009;29(23):7619-7624.
PubMed   |  Link to Article
Salvador  R, Suckling  J, Schwarzbauer  C, Bullmore  E.  Undirected graphs of frequency-dependent functional connectivity in whole brain networks. Philos Trans R Soc Lond B Biol Sci. 2005;360(1457):937-946.
PubMed   |  Link to Article
Sporns  O.  The human connectome: a complex network. Ann N Y Acad Sci.2011;1224:109-125.
PubMed
Sporns  O, Tononi  G, Kötter  R.  The human connectome: a structural description of the human brain. PLoS Comput Biol. 2005;1(4):e42.
PubMed   |  Link to Article
Li  Y, Liu  Y, Li  J,  et al.  Brain anatomical network and intelligence. PLoS Comput Biol. 2009;5(5):e1000395.
PubMed   |  Link to Article
Bassett  DS, Bullmore  ET, Meyer-Lindenberg  A, Apud  JA, Weinberger  DR, Coppola  R.  Cognitive fitness of cost-efficient brain functional networks. Proc Natl Acad Sci U S A. 2009;106(28):11747-11752.
PubMed   |  Link to Article
Fornito  A, Zalesky  A, Pantelis  C, Bullmore  ET.  Schizophrenia, neuroimaging and connectomics. Neuroimage. 2012;62(4):2296-2314.
PubMed   |  Link to Article
van den Heuvel  MP, Kahn  RS.  Abnormal brain wiring as a pathogenetic mechanism in schizophrenia. Biol Psychiatry. 2011;70(12):1107-1108.
PubMed   |  Link to Article
Stephan  KE, Baldeweg  T, Friston  KJ.  Synaptic plasticity and dysconnection in schizophrenia. Biol Psychiatry. 2006;59(10):929-939.
PubMed   |  Link to Article
Bleuler  E. Dementia Praecox or the Group of Schizophrenias. New York, NY: International Universities Press; 1911.
Kanaan  RA, Kim  JS, Kaufmann  WE, Pearlson  GD, Barker  GJ, McGuire  PK.  Diffusion tensor imaging in schizophrenia. Biol Psychiatry. 2005;58(12):921-929.
PubMed   |  Link to Article
Rosenberger  G, Kubicki  M, Nestor  PG,  et al.  Age-related deficits in fronto-temporal connections in schizophrenia: a diffusion tensor imaging study. Schizophr Res. 2008;102(1-3):181-188.
PubMed   |  Link to Article
Voineskos  AN, Lobaugh  NJ, Bouix  S,  et al.  Diffusion tensor tractography findings in schizophrenia across the adult lifespan. Brain. 2010;133(pt 5):1494-1504.
PubMed   |  Link to Article
Ellison-Wright  I, Bullmore  E.  Meta-analysis of diffusion tensor imaging studies in schizophrenia. Schizophr Res. 2009;108(1-3):3-10.
PubMed   |  Link to Article
Nestor  PG, Kubicki  M, Spencer  KM, Niznikiewicz  M, McCarley  RW, Shenton  ME.  Attentional networks and cingulum bundle in chronic schizophrenia. Schizophr Res. 2007;90(1-3):308-315.
PubMed   |  Link to Article
Kubicki  M, Park  H, Westin  CF,  et al.  DTI and MTR abnormalities in schizophrenia: analysis of white matter integrity. Neuroimage. 2005;26(4):1109-1118.
PubMed   |  Link to Article
Fujiwara  H, Namiki  C, Hirao  K,  et al.  Anterior and posterior cingulum abnormalities and their association with psychopathology in schizophrenia: a diffusion tensor imaging study. Schizophr Res. 2007;95(1-3):215-222.
PubMed   |  Link to Article
Whitfield-Gabrieli  S, Thermenos  HW, Milanovic  S,  et al.  Hyperactivity and hyperconnectivity of the default network in schizophrenia and in first-degree relatives of persons with schizophrenia. Proc Natl Acad Sci U S A. 2009;106(4):1279-1284.
PubMed   |  Link to Article
Mannell  MV, Franco  AR, Calhoun  VD, Canive  JM, Thoma  RJ, Mayer  AR.  Resting state and task-induced deactivation: a methodological comparison in patients with schizophrenia and healthy controls. Hum Brain Mapp. 2009;31(3):424-437.
PubMed
Broyd  SJ, Demanuele  C, Debener  S, Helps  SK, James  CJ, Sonuga-Barke  EJ.  Default-mode brain dysfunction in mental disorders: a systematic review. Neurosci Biobehav Rev. 2009;33(3):279-296.
PubMed   |  Link to Article
Bluhm  RL, Miller  J, Lanius  RA,  et al.  Spontaneous low-frequency fluctuations in the BOLD signal in schizophrenic patients: anomalies in the default network. Schizophr Bull. 2007;33(4):1004-1012.
PubMed   |  Link to Article
Williamson  P.  Are anticorrelated networks in the brain relevant to schizophrenia? Schizophr Bull. 2007;33(4):994-1003.
PubMed   |  Link to Article
Garrity  AG, Pearlson  GD, McKiernan  K, Lloyd  D, Kiehl  KA, Calhoun  VD.  Aberrant “default mode” functional connectivity in schizophrenia. Am J Psychiatry. 2007;164(3):450-457.
PubMed   |  Link to Article
Sporns  O, Honey  CJ, Kötter  R.  Identification and classification of hubs in brain networks. PLoS One. 2007;2(10):e1049.
PubMed   |  Link to Article
Tomasi  D, Volkow  ND.  Functional connectivity density mapping. Proc Natl Acad Sci U S A. 2010;107(21):9885-9890.
PubMed   |  Link to Article
Várkuti  B, Cavusoglu  M, Kullik  A,  et al.  Quantifying the link between anatomical connectivity, gray matter volume and regional cerebral blood flow: an integrative MRI study. PLoS One. 2011;6(4):e14801.
PubMed   |  Link to Article
van den Heuvel  MP, Mandl  RCW, Stam  CJ, Kahn  RS, Hulshoff Pol  HE.  Aberrant frontal and temperal network structure in schizophrenia: a graph theoretical analysis. J Neurosci. 2010;30(47):15915-15926.
PubMed   |  Link to Article
van den Heuvel  MP, Stam  CJ, Boersma  M, Hulshoff Pol  HE.  Small-world and scale-free organization of voxel-based resting-state functional connectivity in the human brain. Neuroimage. 2008;43(3):528-539.
PubMed   |  Link to Article
Skudlarski  P, Jagannathan  K, Anderson  K,  et al.  Brain connectivity is not only lower but different in schizophrenia: a combined anatomical and functional approach. Biol Psychiatry. 2010;68(1):61-69.
PubMed   |  Link to Article
Lynall  ME, Bassett  DS, Kerwin  R,  et al.  Functional connectivity and brain networks in schizophrenia. J Neurosci. 2010;30(28):9477-9487.
PubMed   |  Link to Article
Liu  Y, Liang  M, Zhou  Y,  et al.  Disrupted small-world networks in schizophrenia. Brain. 2008;131(pt 4):945-961.
PubMed   |  Link to Article
Zalesky  A, Fornito  A, Harding  IH,  et al.  Whole-brain anatomical networks: does the choice of nodes matter? Neuroimage. 2010;50(3):970-983.
PubMed   |  Link to Article
Rubinov  M, Bassett  DS.  Emerging evidence of connectomic abnormalities in schizophrenia. J Neurosci. 2011;31(17):6263-6265.
PubMed   |  Link to Article
van den Heuvel  MP, Sporns  O.  Rich-club organization of the human connectome. J Neurosci. 2011;31(44):15775-15786.
PubMed   |  Link to Article
Zamora-Lopez  G, Zhou  C, Kurths  J.  Cortical hubs form a module for multisensory integration on top of the hierarchy of cortical networks. Front Neuroinform. 2010;4:1.
PubMed
van den Heuvel  MP, Kahn  RS, Goni  J, Sporns  O.  A high cost, high capacity backbone for global brain communication. Proc Natl Acad Sci U S A. 2012;109(28):11372-11377.
PubMed   |  Link to Article
van den Heuvel  MP, Mandl  RC, Hulshoff Pol  HE.  Normalized group clustering of resting-state fMRI data. PLoS One. 2008;3(4):e2001.
PubMed   |  Link to Article
van den Heuvel  MP, Mandl  RCW, Kahn  RS, Hulshoff Pol  HE.  Functionally linked resting-state networks reflect the underlying structural connectivity architecture of the human brain. Hum Brain Mapp. 2009;30(10):3127-3141.
PubMed   |  Link to Article
Mandl  RC, Schnack  HG, Luigjes  J,  et al.  Tract-based analysis of magnetization transfer ratio and diffusion tensor imaging of the frontal and frontotemporal connections in schizophrenia. Schizophr Bull. 2008;36(4):778-787.
PubMed   |  Link to Article
Collin  G, Hulshoff Pol  HE, Haijma  SV, Cahn  W, Kahn  RS, van den Heuvel  MP.  Impaired cerebellar functional connectivity in schizophrenia patients and their healthy siblings. Front Psychiatry. 2011;2:73.
PubMed   |  Link to Article
Fischel  B, van der Kouwe  A, Destrieux  C,  et al.  Automatically parcellating the human cerebral cortex. Cereb Cortex. 2004;14(1):11-22.
PubMed   |  Link to Article
Van Dijk  KR, Sabuncu  MR, Buckner  RL.  The influence of head motion on intrinsic functional connectivity MRI. Neuroimage. 2012;59(1):431-438.
PubMed   |  Link to Article
Power  JD, Barnes  KA, Snyder  AZ, Schlaggar  BL, Petersen  SE.  Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage. 2012;59(3):2142-2154.
PubMed   |  Link to Article
Fox  MD, Zhang  D, Snyder  AZ, Raichle  ME.  The global signal and observed anticorrelated resting state brain networks. J Neurophysiol. 2009;101(6):3270-3283.
PubMed   |  Link to Article
Saad  ZS, Gotts  SJ, Murphy  K,  et al.  Trouble at rest: how correlation patterns and group differences become distorted after global signal regression. Brain Connectivity. 2012;2(1):25-32.
PubMed   |  Link to Article
Rubinov  M, Sporns  O.  Complex network measures of brain connectivity: uses and interpretations. Neuroimage. 2010;52(3):1059-1069.
PubMed   |  Link to Article
Newman  ME.  Modularity and community structure in networks. Proc Natl Acad Sci U S A. 2006;103(23):8577-8582.
PubMed   |  Link to Article
Colizza  V, Flammini  A, Serrano  MA, Vespignani  A.  Detecting rich-club ordering in complex networks. Nat Phys. 2006;2:110-115.
Link to Article
Opsahl  T, Colizza  V, Panzarasa  P, Ramasco  JJ.  Prominence and control: the weighted rich-club effect. Phys Rev Lett. 2008;101(16):168702.
PubMed   |  Link to Article
McAuley  JJ, da Fontoura Costa  L, Caetano  TS.  Rich-club phenomena across complex network hierachies. Appl Phys Lett. 2007;91:084103.
Link to Article
de Reus  MA, van den Heuvel  MP.  Estimating false positives and negatives in brain networks. Neuroimage. 2013;70:402-409.
PubMed   |  Link to Article
Honey  CJ, Sporns  O, Cammoun  L,  et al.  Predicting human resting-state functional connectivity from structural connectivity. Proc Natl Acad Sci U S A. 2009;106(6):2035-2040.
PubMed   |  Link to Article
Bassett  DS, Bullmore  E, Verchinski  BA, Mattay  VS, Weinberger  DR, Meyer-Lindenberg  A.  Hierarchical organization of human cortical networks in health and schizophrenia. J Neurosci. 2008;28(37):9239-9248.
PubMed   |  Link to Article
Zalesky  A, Fornito  A, Seal  ML,  et al.  Disrupted axonal fiber connectivity in schizophrenia. Biol Psychiatry. 2011;69(1):80-90.
PubMed   |  Link to Article
Zuo  XN, Ehmke  R, Mennes  M,  et al.  Network centrality in the human functional connectome. Cereb Cortex. 2012;22(8):1862-1875.
PubMed   |  Link to Article
Kubicki  M, McCarley  R, Westin  CF,  et al.  A review of diffusion tensor imaging studies in schizophrenia. J Psychiatr Res. 2007;41(1-2):15-30.
PubMed   |  Link to Article
Zhang  Y, Schuff  N, Jahng  GH,  et al.  Diffusion tensor imaging of cingulum fibers in mild cognitive impairment and Alzheimer disease. Neurology. 2007;68(1):13-19.
PubMed   |  Link to Article
Harringer  L, van den Heuvel  MP, Sporns  O.  Rich club organization of macaque cerebral cortex and its role in network communication. PLoS ONE. 2012;7(9):e46497.
PubMed   |  Link to Article
Kubicki  M, Westin  CF, Maier  SE,  et al.  Uncinate fasciculus findings in schizophrenia: a magnetic resonance diffusion tensor imaging study. Am J Psychiatry. 2002;159(5):813-820.
PubMed   |  Link to Article
Zamora-Lopez  G, Zhou  C, Kurths  J.  Exploring brain function from anatomical connectivity. Front Neurosci. 2011;5:83.
PubMed   |  Link to Article
Cabral  J, Hugues  E, Kringelbach  ML, Deco  G.  Modeling the outcome of structural disconnection on resting-state functional connectivity. Neuroimage. 2012;62(3):1342-1353.
PubMed   |  Link to Article
Cammoun  L, Gigandet  X, Meskaldji  D,  et al.  Mapping the human connectome at multiple scales with diffusion spectrum MRI. J Neurosci Methods. 2012;203(2):386-397.
PubMed   |  Link to Article
Bassett  DS, Brown  JA, Deshpande  V, Carlson  JM, Grafton  ST.  Conserved and variable architecture of human white matter connectivity. Neuroimage. 2011;54(2):1262-1279.
PubMed   |  Link to Article

Figures

Place holder to copy figure label and caption
Figure 1.
The Analysis Steps

A, Individual T1 images were used for automatic classification of gray and white matter tissue and parcellation of the cortex into 68 distinct brain regions, forming the nodes of the individual brain networks. B, Streamline tractography was applied to the diffusion tension imaging (DTI) data to reconstruct cortico-cortical white matter pathways. From the set of reconstructed streamlines, streamlines that interconnected region i and j from the set of 68 or 82 regions were taken as an edge between node i and j in the structural brain network. Streamline count was taken to represent the weight of the connection and was aggregated into a structural connectivity (SC) matrix. C, Functional connectivity (FC) between node i and j was computed as their level of correlation between their resting-state function magnetic resonance imaging (fMRI) blood oxygenation level dependent (BOLD) time series, resulting in a matrix FC. D, The topologic organization of the resulting individual structural brain networks was examined, including (among other metrics) measurements of the rich club, global strength, and global efficiency (top). The level of coupling between SC and FC was examined by computing the level of correlation between the weights of (existing) structural connections and their functional counterparts. This correlation is referred to as the level of SC-FC coupling. At the group level, values were examined between the patient and control groups (for both the principal and replication data sets). Statistical evaluation was performed using permutation testing (see Methods and eMethods in Supplement).

Graphic Jump Location
Place holder to copy figure label and caption
Figure 2.
Rich Club Organization

A, Group-averaged rich club curve for controls (white) and patients (blue). Patients had a significantly reduced rich club organization for the range k = 26 to k = 28, reflecting a lower level of connectivity between central hubs of the brain (cortical: P = .003; whole brain: P = .04; 10 000 permutations) B, Confirming previous findings, rich club members included the bilateral precuneus, superior frontal cortex, superior parietal cortex, and the insula in both the healthy and patient populations. This figure is based on the group-averaged cortical network in controls (at a rich club level of k >15).53 C, Edges across individual brain networks (both controls and patients) were divided into 3 distinct classes: rich club connections linking rich club members (red), feeder connections linking rich club members to non–rich club members (orange), and local connections connecting non–rich club members (yellow edges). Examining the density of rich club, feeder, and local connections between the populations of controls and patients revealed a significant reduction in rich club density in patients but no significant effect in density of feeder and local connections. The figure shows mean (SD) density values for each of the 3 classes, scaled to the mean density values of the control group. ROI indicates region of interest.

Graphic Jump Location
Place holder to copy figure label and caption
Figure 3.
Replication Data Set

A, Group-averaged rich club curve of controls (white) and patients (blue) of the replication data set (n = 41 patients and 51 controls). B, Global graph metrics (cortex network) of controls (dark blue) and patients (light blue). No differences were found between patients and controls on connectivity strength (S), global efficiency (GE), or clustering (C). mod indicates modularity. C, Density of rich club, feeder, and local connections. Confirming the findings of the principal data set, patients had a significant reduction in rich club density (*P < .05). D, Structural connectivity (SC)–functional connectivity (FC) coupling for patients and controls (data of 39 patients and 35 controls). Confirming the findings of the principal data set, patients had an increased level of SC-FC coupling.

Graphic Jump Location
Place holder to copy figure label and caption
Figure 4.
Global Efficiency and Rich Club Density

A, General topologic graph metrics. Patients had reduced levels of connectivity strength (S) and global efficiency (GE) and increased levels of local clustering (C) and modularity (mod) (*P < .05, permutation testing, 10 000 permutations, effects of GE remained significant after volume and global S correction, cortex network). B, Across the control (top) and patient (bottom) populations, global efficiency (y-axis) was significantly correlated with rich club density (x-axis), after correcting for overall differences in global connectivity S.

Graphic Jump Location
Place holder to copy figure label and caption
Figure 5.
Structural Connectivity (SC)–Functional Connectivity (FC) Coupling

A, Patients revealed a significant increase in SC-FC coupling compared with healthy controls (*P < .05, permutation testing, 10 000 permutations). B, In the patient population (bottom), lower levels of normalized rich club density (measured as the fraction of streamlines inside the rich club relative to the total number of streamlines, corrected for volume effects, blue) were associated with increased SC-FC coupling. This association was absent in the control population (top).

Graphic Jump Location

Tables

Table Graphic Jump LocationTable.  Demographic and Clinical Characteristics of Principal and Replication Data Sets

References

Hagmann  P, Cammoun  L, Gigandet  X,  et al.  Mapping the structural core of human cerebral cortex. PLoS Biol. 2008;6(7):e159.
PubMed   |  Link to Article
Sporns  O, Chialvo  DR, Kaiser  M, Hilgetag  CC.  Organization, development and function of complex brain networks. Trends Cogn Sci. 2004;8(9):418-425.
PubMed   |  Link to Article
van den Heuvel  MP, Stam  CJ, Kahn  RS, Hulshoff Pol  HE.  Efficiency of functional brain networks and intellectual performance. J Neurosci. 2009;29(23):7619-7624.
PubMed   |  Link to Article
Salvador  R, Suckling  J, Schwarzbauer  C, Bullmore  E.  Undirected graphs of frequency-dependent functional connectivity in whole brain networks. Philos Trans R Soc Lond B Biol Sci. 2005;360(1457):937-946.
PubMed   |  Link to Article
Sporns  O.  The human connectome: a complex network. Ann N Y Acad Sci.2011;1224:109-125.
PubMed
Sporns  O, Tononi  G, Kötter  R.  The human connectome: a structural description of the human brain. PLoS Comput Biol. 2005;1(4):e42.
PubMed   |  Link to Article
Li  Y, Liu  Y, Li  J,  et al.  Brain anatomical network and intelligence. PLoS Comput Biol. 2009;5(5):e1000395.
PubMed   |  Link to Article
Bassett  DS, Bullmore  ET, Meyer-Lindenberg  A, Apud  JA, Weinberger  DR, Coppola  R.  Cognitive fitness of cost-efficient brain functional networks. Proc Natl Acad Sci U S A. 2009;106(28):11747-11752.
PubMed   |  Link to Article
Fornito  A, Zalesky  A, Pantelis  C, Bullmore  ET.  Schizophrenia, neuroimaging and connectomics. Neuroimage. 2012;62(4):2296-2314.
PubMed   |  Link to Article
van den Heuvel  MP, Kahn  RS.  Abnormal brain wiring as a pathogenetic mechanism in schizophrenia. Biol Psychiatry. 2011;70(12):1107-1108.
PubMed   |  Link to Article
Stephan  KE, Baldeweg  T, Friston  KJ.  Synaptic plasticity and dysconnection in schizophrenia. Biol Psychiatry. 2006;59(10):929-939.
PubMed   |  Link to Article
Bleuler  E. Dementia Praecox or the Group of Schizophrenias. New York, NY: International Universities Press; 1911.
Kanaan  RA, Kim  JS, Kaufmann  WE, Pearlson  GD, Barker  GJ, McGuire  PK.  Diffusion tensor imaging in schizophrenia. Biol Psychiatry. 2005;58(12):921-929.
PubMed   |  Link to Article
Rosenberger  G, Kubicki  M, Nestor  PG,  et al.  Age-related deficits in fronto-temporal connections in schizophrenia: a diffusion tensor imaging study. Schizophr Res. 2008;102(1-3):181-188.
PubMed   |  Link to Article
Voineskos  AN, Lobaugh  NJ, Bouix  S,  et al.  Diffusion tensor tractography findings in schizophrenia across the adult lifespan. Brain. 2010;133(pt 5):1494-1504.
PubMed   |  Link to Article
Ellison-Wright  I, Bullmore  E.  Meta-analysis of diffusion tensor imaging studies in schizophrenia. Schizophr Res. 2009;108(1-3):3-10.
PubMed   |  Link to Article
Nestor  PG, Kubicki  M, Spencer  KM, Niznikiewicz  M, McCarley  RW, Shenton  ME.  Attentional networks and cingulum bundle in chronic schizophrenia. Schizophr Res. 2007;90(1-3):308-315.
PubMed   |  Link to Article
Kubicki  M, Park  H, Westin  CF,  et al.  DTI and MTR abnormalities in schizophrenia: analysis of white matter integrity. Neuroimage. 2005;26(4):1109-1118.
PubMed   |  Link to Article
Fujiwara  H, Namiki  C, Hirao  K,  et al.  Anterior and posterior cingulum abnormalities and their association with psychopathology in schizophrenia: a diffusion tensor imaging study. Schizophr Res. 2007;95(1-3):215-222.
PubMed   |  Link to Article
Whitfield-Gabrieli  S, Thermenos  HW, Milanovic  S,  et al.  Hyperactivity and hyperconnectivity of the default network in schizophrenia and in first-degree relatives of persons with schizophrenia. Proc Natl Acad Sci U S A. 2009;106(4):1279-1284.
PubMed   |  Link to Article
Mannell  MV, Franco  AR, Calhoun  VD, Canive  JM, Thoma  RJ, Mayer  AR.  Resting state and task-induced deactivation: a methodological comparison in patients with schizophrenia and healthy controls. Hum Brain Mapp. 2009;31(3):424-437.
PubMed
Broyd  SJ, Demanuele  C, Debener  S, Helps  SK, James  CJ, Sonuga-Barke  EJ.  Default-mode brain dysfunction in mental disorders: a systematic review. Neurosci Biobehav Rev. 2009;33(3):279-296.
PubMed   |  Link to Article
Bluhm  RL, Miller  J, Lanius  RA,  et al.  Spontaneous low-frequency fluctuations in the BOLD signal in schizophrenic patients: anomalies in the default network. Schizophr Bull. 2007;33(4):1004-1012.
PubMed   |  Link to Article
Williamson  P.  Are anticorrelated networks in the brain relevant to schizophrenia? Schizophr Bull. 2007;33(4):994-1003.
PubMed   |  Link to Article
Garrity  AG, Pearlson  GD, McKiernan  K, Lloyd  D, Kiehl  KA, Calhoun  VD.  Aberrant “default mode” functional connectivity in schizophrenia. Am J Psychiatry. 2007;164(3):450-457.
PubMed   |  Link to Article
Sporns  O, Honey  CJ, Kötter  R.  Identification and classification of hubs in brain networks. PLoS One. 2007;2(10):e1049.
PubMed   |  Link to Article
Tomasi  D, Volkow  ND.  Functional connectivity density mapping. Proc Natl Acad Sci U S A. 2010;107(21):9885-9890.
PubMed   |  Link to Article
Várkuti  B, Cavusoglu  M, Kullik  A,  et al.  Quantifying the link between anatomical connectivity, gray matter volume and regional cerebral blood flow: an integrative MRI study. PLoS One. 2011;6(4):e14801.
PubMed   |  Link to Article
van den Heuvel  MP, Mandl  RCW, Stam  CJ, Kahn  RS, Hulshoff Pol  HE.  Aberrant frontal and temperal network structure in schizophrenia: a graph theoretical analysis. J Neurosci. 2010;30(47):15915-15926.
PubMed   |  Link to Article
van den Heuvel  MP, Stam  CJ, Boersma  M, Hulshoff Pol  HE.  Small-world and scale-free organization of voxel-based resting-state functional connectivity in the human brain. Neuroimage. 2008;43(3):528-539.
PubMed   |  Link to Article
Skudlarski  P, Jagannathan  K, Anderson  K,  et al.  Brain connectivity is not only lower but different in schizophrenia: a combined anatomical and functional approach. Biol Psychiatry. 2010;68(1):61-69.
PubMed   |  Link to Article
Lynall  ME, Bassett  DS, Kerwin  R,  et al.  Functional connectivity and brain networks in schizophrenia. J Neurosci. 2010;30(28):9477-9487.
PubMed   |  Link to Article
Liu  Y, Liang  M, Zhou  Y,  et al.  Disrupted small-world networks in schizophrenia. Brain. 2008;131(pt 4):945-961.
PubMed   |  Link to Article
Zalesky  A, Fornito  A, Harding  IH,  et al.  Whole-brain anatomical networks: does the choice of nodes matter? Neuroimage. 2010;50(3):970-983.
PubMed   |  Link to Article
Rubinov  M, Bassett  DS.  Emerging evidence of connectomic abnormalities in schizophrenia. J Neurosci. 2011;31(17):6263-6265.
PubMed   |  Link to Article
van den Heuvel  MP, Sporns  O.  Rich-club organization of the human connectome. J Neurosci. 2011;31(44):15775-15786.
PubMed   |  Link to Article
Zamora-Lopez  G, Zhou  C, Kurths  J.  Cortical hubs form a module for multisensory integration on top of the hierarchy of cortical networks. Front Neuroinform. 2010;4:1.
PubMed
van den Heuvel  MP, Kahn  RS, Goni  J, Sporns  O.  A high cost, high capacity backbone for global brain communication. Proc Natl Acad Sci U S A. 2012;109(28):11372-11377.
PubMed   |  Link to Article
van den Heuvel  MP, Mandl  RC, Hulshoff Pol  HE.  Normalized group clustering of resting-state fMRI data. PLoS One. 2008;3(4):e2001.
PubMed   |  Link to Article
van den Heuvel  MP, Mandl  RCW, Kahn  RS, Hulshoff Pol  HE.  Functionally linked resting-state networks reflect the underlying structural connectivity architecture of the human brain. Hum Brain Mapp. 2009;30(10):3127-3141.
PubMed   |  Link to Article
Mandl  RC, Schnack  HG, Luigjes  J,  et al.  Tract-based analysis of magnetization transfer ratio and diffusion tensor imaging of the frontal and frontotemporal connections in schizophrenia. Schizophr Bull. 2008;36(4):778-787.
PubMed   |  Link to Article
Collin  G, Hulshoff Pol  HE, Haijma  SV, Cahn  W, Kahn  RS, van den Heuvel  MP.  Impaired cerebellar functional connectivity in schizophrenia patients and their healthy siblings. Front Psychiatry. 2011;2:73.
PubMed   |  Link to Article
Fischel  B, van der Kouwe  A, Destrieux  C,  et al.  Automatically parcellating the human cerebral cortex. Cereb Cortex. 2004;14(1):11-22.
PubMed   |  Link to Article
Van Dijk  KR, Sabuncu  MR, Buckner  RL.  The influence of head motion on intrinsic functional connectivity MRI. Neuroimage. 2012;59(1):431-438.
PubMed   |  Link to Article
Power  JD, Barnes  KA, Snyder  AZ, Schlaggar  BL, Petersen  SE.  Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage. 2012;59(3):2142-2154.
PubMed   |  Link to Article
Fox  MD, Zhang  D, Snyder  AZ, Raichle  ME.  The global signal and observed anticorrelated resting state brain networks. J Neurophysiol. 2009;101(6):3270-3283.
PubMed   |  Link to Article
Saad  ZS, Gotts  SJ, Murphy  K,  et al.  Trouble at rest: how correlation patterns and group differences become distorted after global signal regression. Brain Connectivity. 2012;2(1):25-32.
PubMed   |  Link to Article
Rubinov  M, Sporns  O.  Complex network measures of brain connectivity: uses and interpretations. Neuroimage. 2010;52(3):1059-1069.
PubMed   |  Link to Article
Newman  ME.  Modularity and community structure in networks. Proc Natl Acad Sci U S A. 2006;103(23):8577-8582.
PubMed   |  Link to Article
Colizza  V, Flammini  A, Serrano  MA, Vespignani  A.  Detecting rich-club ordering in complex networks. Nat Phys. 2006;2:110-115.
Link to Article
Opsahl  T, Colizza  V, Panzarasa  P, Ramasco  JJ.  Prominence and control: the weighted rich-club effect. Phys Rev Lett. 2008;101(16):168702.
PubMed   |  Link to Article
McAuley  JJ, da Fontoura Costa  L, Caetano  TS.  Rich-club phenomena across complex network hierachies. Appl Phys Lett. 2007;91:084103.
Link to Article
de Reus  MA, van den Heuvel  MP.  Estimating false positives and negatives in brain networks. Neuroimage. 2013;70:402-409.
PubMed   |  Link to Article
Honey  CJ, Sporns  O, Cammoun  L,  et al.  Predicting human resting-state functional connectivity from structural connectivity. Proc Natl Acad Sci U S A. 2009;106(6):2035-2040.
PubMed   |  Link to Article
Bassett  DS, Bullmore  E, Verchinski  BA, Mattay  VS, Weinberger  DR, Meyer-Lindenberg  A.  Hierarchical organization of human cortical networks in health and schizophrenia. J Neurosci. 2008;28(37):9239-9248.
PubMed   |  Link to Article
Zalesky  A, Fornito  A, Seal  ML,  et al.  Disrupted axonal fiber connectivity in schizophrenia. Biol Psychiatry. 2011;69(1):80-90.
PubMed   |  Link to Article
Zuo  XN, Ehmke  R, Mennes  M,  et al.  Network centrality in the human functional connectome. Cereb Cortex. 2012;22(8):1862-1875.
PubMed   |  Link to Article
Kubicki  M, McCarley  R, Westin  CF,  et al.  A review of diffusion tensor imaging studies in schizophrenia. J Psychiatr Res. 2007;41(1-2):15-30.
PubMed   |  Link to Article
Zhang  Y, Schuff  N, Jahng  GH,  et al.  Diffusion tensor imaging of cingulum fibers in mild cognitive impairment and Alzheimer disease. Neurology. 2007;68(1):13-19.
PubMed   |  Link to Article
Harringer  L, van den Heuvel  MP, Sporns  O.  Rich club organization of macaque cerebral cortex and its role in network communication. PLoS ONE. 2012;7(9):e46497.
PubMed   |  Link to Article
Kubicki  M, Westin  CF, Maier  SE,  et al.  Uncinate fasciculus findings in schizophrenia: a magnetic resonance diffusion tensor imaging study. Am J Psychiatry. 2002;159(5):813-820.
PubMed   |  Link to Article
Zamora-Lopez  G, Zhou  C, Kurths  J.  Exploring brain function from anatomical connectivity. Front Neurosci. 2011;5:83.
PubMed   |  Link to Article
Cabral  J, Hugues  E, Kringelbach  ML, Deco  G.  Modeling the outcome of structural disconnection on resting-state functional connectivity. Neuroimage. 2012;62(3):1342-1353.
PubMed   |  Link to Article
Cammoun  L, Gigandet  X, Meskaldji  D,  et al.  Mapping the human connectome at multiple scales with diffusion spectrum MRI. J Neurosci Methods. 2012;203(2):386-397.
PubMed   |  Link to Article
Bassett  DS, Brown  JA, Deshpande  V, Carlson  JM, Grafton  ST.  Conserved and variable architecture of human white matter connectivity. Neuroimage. 2011;54(2):1262-1279.
PubMed   |  Link to Article

Correspondence

CME


You need to register in order to view this quiz.
Submit a Comment

Multimedia

Supplement.

eMethods

eResults

eReferences

eFigure 1. Rich Club Connectivity Density Effects at Different Rich Club Levels

eFigure 2. Connectivity Density of Rich Club, Feeder and Local Connections With the Rich Club Defined on an Individual Basis

eFigure 3. Relationship Between Regional Volume and Connectivity

Supplemental Content

Some tools below are only available to our subscribers or users with an online account.

Web of Science® Times Cited: 16

Related Content

Customize your page view by dragging & repositioning the boxes below.

See Also...
Articles Related By Topic
Related Collections
JAMAevidence.com

Users' Guides to the Medical Literature
Clinical Resolution

Users' Guides to the Medical Literature
Clinical Scenario