Skip to main content

Main menu

  • HOME
  • CONTENT
    • Early Release
    • Featured
    • Current Issue
    • Issue Archive
    • Collections
    • Podcast
  • ALERTS
  • FOR AUTHORS
    • Information for Authors
    • Fees
    • Journal Clubs
    • eLetters
    • Submit
    • Special Collections
  • EDITORIAL BOARD
    • Editorial Board
    • ECR Advisory Board
    • Journal Staff
  • ABOUT
    • Overview
    • Advertise
    • For the Media
    • Rights and Permissions
    • Privacy Policy
    • Feedback
    • Accessibility
  • SUBSCRIBE

User menu

  • Log out
  • Log in
  • My Cart

Search

  • Advanced search
Journal of Neuroscience
  • Log out
  • Log in
  • My Cart
Journal of Neuroscience

Advanced Search

Submit a Manuscript
  • HOME
  • CONTENT
    • Early Release
    • Featured
    • Current Issue
    • Issue Archive
    • Collections
    • Podcast
  • ALERTS
  • FOR AUTHORS
    • Information for Authors
    • Fees
    • Journal Clubs
    • eLetters
    • Submit
    • Special Collections
  • EDITORIAL BOARD
    • Editorial Board
    • ECR Advisory Board
    • Journal Staff
  • ABOUT
    • Overview
    • Advertise
    • For the Media
    • Rights and Permissions
    • Privacy Policy
    • Feedback
    • Accessibility
  • SUBSCRIBE
PreviousNext
Research Articles, Systems/Circuits

Inner Hemispheric and Interhemispheric Connectivity Balance in the Human Brain

Ronnie Krupnik, Yossi Yovel and Yaniv Assaf
Journal of Neuroscience 6 October 2021, 41 (40) 8351-8361; https://doi.org/10.1523/JNEUROSCI.1074-21.2021
Ronnie Krupnik
1Sagol School of Neuroscience, Tel Aviv University, Tel Aviv 69978, Israel
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
Yossi Yovel
1Sagol School of Neuroscience, Tel Aviv University, Tel Aviv 69978, Israel
2School of Zoology, George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv 69978, Israel
3Steinhardt Museum of Natural History, Tel Aviv University, Tel Aviv 69978, Israel
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
Yaniv Assaf
1Sagol School of Neuroscience, Tel Aviv University, Tel Aviv 69978, Israel
4School of Neurobiology, Biochemistry and Biophysics, George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv 69978, Israel
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
  • Article
  • Figures & Data
  • Info & Metrics
  • eLetters
  • PDF
Loading

Abstract

The connectome of the brain has a great impact on the function of the brain as the structure of the connectome affects the speed and efficiency of information transfer. As a highly energy-consuming organ, an efficient network structure is essential. A previous study has shown consistent overall brain connectivity across a large variety of species. This connectivity conservation was explained by a balance between interhemispheric and intrahemispheric connections; that is, spices with highly connected hemispheres appear to have weaker interhemisphere connections. This study examines this connectivity trade-off in the human brain using diffusion-based tractography and network analysis in the Human Connectome Project (970 subjects, 527 female). We explore the biological origins of this phenomenon, heritability, and the effect on cognitive measures.

The proportion of commissural fibers in the brain had a negative correlation to hemispheric efficiency, pointing to a trade-off between inner hemispheric and interhemispheric connectivity. Network hubs including anterior and middle cingulate cortex, superior frontal areas, medial occipital areas, the parahippocampal gyrus, post- and precentral gyri, and the precuneus had the strongest contribution to this phenomenon. Other results show a high heritability as well as a strong connection to crystalized intelligence. This work presents cohort-based network analysis research, spanning a large variety of samples and exploring the overall architecture of the human connectome. Our results show a connectivity conservation phenomenon at the base of the overall brain network architecture. This network structure may explain much of the functional, behavioral, and cognitive variability among different brains.

SIGNIFICANCE STATEMENT The network structure of the brain is at the basis of every brain function as it dictates the characteristics of information transfer. Understanding the patterns and mechanisms that guide the connectome structure is crucial to understanding the brain itself. Here we unravel the mechanism at the base of the connectivity conservation phenomenon by exploring the interaction between hemispheric and commissural connectivity in a large-scale cohort-based connectivity study. We describe the trade-off between the two components and examine the origins of the trade-off and observe the effect on cognitive abilities and behavior.

  • connectivity
  • imaging
  • MRI
  • network
  • tractography

Introduction

The white matter of the brain can, quite intuitively, be modeled as edges in a network as it constitutes the connections among different nodes in the brain (Bullmore and Bassett, 2011). This network, the connectome, is assumed to have a great impact on brain functionality as it dictates the paths of information transfer in the brain (Bullmore and Sporns, 2012; Mišić et al., 2014).

An earlier study showed that the brain network connectivity is balanced and conserved across the brains of the mammalian class (Assaf et al., 2020). That study, exploring a large variety of mammalian brains, revealed that the overall connectivity, as described by the characteristic path length, is consistent across species, despite the large variance in brain volume and network structure. This connectivity consistency was manifested by a balance between interhemispheric and intrahemispheric connections. That is, the brain networks of species with fewer commissural connections seemed to compensate with shorter paths between node pairs, pointing to a more efficient network. This trade-off was also observed in human subjects with agenesis of the corpus callosum, presenting a strong hemispheric network, thus suggesting a mechanism that compensates for the lack of commissural connectivity (Owen et al., 2013a,b). This inherent balance among the networks may explain the connectivity conservation phenomenon across species.

In this study, we examined the connectivity conservation phenomenon in a large database of human subjects. Although we must take into account the interspecies variability that is significantly larger than interhuman variability, functional studies show that different subject connectomes have different and individually unique connectivity patterns (Finn et al., 2015). Thus different human subjects are still expected to show a variety of network structures while conserving the overall network connectivity, independently of brain size or the number of tracts.

Unlike with animals, analyzing this phenomenon in the well-mapped and long-studied human brain allows us to compare network properties at the node level rather than globally. More specifically, by using the same coordinate system for all human subjects, we can pinpoint the anatomic and functional origins of the connectivity conservation, an exploration that could not be done across species.

The brain network has a small-world structure, as was demonstrated by functional (Achard et al., 2006; Achard and Bullmore, 2007; van den Heuvel et al., 2008), structural (He et al., 2007), and diffusion tensor imaging (Sporns et al., 2004). This structure has been linked to many aspects of brain function as it could be affected by disease (Stam et al., 2007; Bassett and Bullmore, 2009; Stam, 2014), aging (Zuo et al., 2017) and pharmacological interventions (Achard and Bullmore, 2007). As network efficiency is highly dependent on this small-world structure (Latora and Marchiori, 2001), it may be expected to have a key role in many brain functions. Brain network efficiency, both structural and functional, was shown to affect cognition and other brain functions as loss of network efficiency in aging or disease was associated with a decline in cognitive abilities (Vlooswijk et al., 2011; Wen et al., 2011; Wiseman et al., 2018), whereas local and global efficiency both predict intelligence (Fischer et al., 2014; Pineda-Pardo et al., 2016). As such, it can be assumed that a trade-off between hemispheric efficiency and commissural connectivity affects cognition. The analysis of connectivity conservation across humans could be used to reveal cognitive implications of this structural phenomenon, leading to a greater understanding of brain organization and the effect on brain function.

In this work, we explore the connectivity conservation phenomenon in a large cohort of human subjects. We use the Human Connectome Project (HCP) young adult dataset (Van Essen et al., 2012) to observe interhemispherc and intrahemispheric connectivity across subjects and examine the possible structural origins for this balance. We used network analysis measures to locate brain regions that contribute to the balance as the more central, rich-club nodes. We also point to a possible hereditary source for this phenomenon, as well as the effect on cognitive abilities.

Materials and Methods

Data

We used diffusion and T1 data for 970 healthy adults (age 22–37 years, mean 28.8, 527 females) from the HCP 1200 young adult release (Van Essen et al., 2012). This includes high-quality T1w structural images as well as multishell diffusion scans. Full protocol details are available in the HCP reference manual (https://www.humanconnectome.org/storage/app/media/documentation/s1200/HCP_S1200_Release_Reference_Manual.pdf).

Personal information for each subject consisted of data from the National Institutes of Health (NIH) Toolbox questionnaires (NIH Toolbox (https://www.healthmeasures.net/explore-measurement-systems/nih-toolbox); Gershon et al., 2010) measuring cognitive abilities. In addition to that, family information was also collected, and genetic samples were taken for some participants. Of those, 105 pairs of subjects were genetically confirmed as monozygotic (MZ) twins, and 65 as dizygotic (DZ) twins, 95 were nontwin siblings and 270 were not related or third siblings of twin pairs. For heritability analysis, we only used data from twin pairs with genetic confirmation.

Experimental design

Preprocessing

The HCP provides minimally preprocessed images (Glasser et al., 2013). This preprocessing pipeline includes intensity normalization across runs, topup and eddy corrections, gradient nonlinear correction, and registration of diffusion image to MPRAGE using the Functional MRI of the Brain (FMRIB) linear image registration tool boundary-based registration (BRB)+bbregister. The full pipeline is available online (https://github.com/Washington-University/Pipelines).

To reduce the potential effect of brain volume, we resampled all images to a similar number of voxels per brain, that is, the resized brain mask for each subject had the same number of voxels as the smallest brain mask in the dataset. Then five tissue types (CSF, cortical gray matter, subcortical gray matter, white matter, and nonbrain tissue) were segmented on the structural image using the FMRIB Software Library (FSL) algorithm on MRtrix3 (Zhang et al., 2001; Jenkinson et al., 2002; Smith et al., 2004, 2012; Patenaude et al., 2011; Tournier et al., 2019).

Cortical parcellation

For anatomic parcellation, we used a modified automated anatomical labeling (AAL) atlas (Tzourio-Mazoyer et al., 2002). Subcortical and cerebellum areas were removed from the original atlas. As connectivity measures such as node degree are statistically biased toward larger areas, we divided the remaining brain regions into 150 similarly sized areas in each hemisphere (Cammoun et al., 2012) using a k means clustering algorithm. These areas were divided so that each new label consists of voxels from a single original AAL label, but the larger an original label was, the more areas it was divided into. This new (AAL300) atlas was registered to each subject's native space using the FMRIB FSL nonlinear registration tool FNIRT.

Diffusion-based tractography

Multishell multitissue constrained spherical deconvolution (CSD) (Tournier et al., 2004, 2019; Jeurissen et al., 2014; Dhollander et al., 2016), followed by anatomically constrained deterministic tractography, was performed using MRtrix3. To eliminate biases because of the large intersubject variability in brain volumes, we normalized the tractography parameters by each subject's brain volume.

Two tractograms were computed. In one, we filtered tracts to include only those that both start and end in or near voxels included in the cortical atlas with a search radius of two voxels to account for registration errors (Yeh et al., 2019). The remaining streamlines were then further filtered using the anatomically constrained spherical-deconvolution informed filtering of tractograms (SIFT) algorithm (Smith et al., 2013) to a set number of streamlines for each subject. On the second tractogram (full brain tracts), we used tracts from the entire brain mask and then filtered them using the SIFT algorithm. Here we removed the corticospinal tracts using a manually drawn “not” mask. This mask was created in Montreal Neurological Institute (MNI) space and registered to each subject's native space using FNIRT.

Network matrices

We computed connectivity matrices for individual brain networks. For each pair of brain regions in the modified AAL300 atlas, a connection was added if any fibers connecting the two areas were found. Streamlines with end points outside the atlas mask were discarded if no atlas area was found within a two voxel radius. We set the distance between each pair of nodes as the reciprocal of normalized number of fibers found between the two nodes, under the assumption that this weighting represents the probability of information transfer between those nodes (Sotiropoulos and Zalesky, 2019). This was done for a random subset of 15,000 tracts of the atlas-based tractogram, over 100 permutations. Our final connectivity matrix consisted of the mean weight of each edge over all permutations.

To compute a group-averaged connectivity matrix, we averaged the connection strength for each pair of nodes across subjects. Connections that appeared in 75% of matrices or less were discarded.

For normalizations and statistical analysis, we generated 1000 random matrices. The common model for random brain networks is the Maslov–Sneppen algorithm (Maslov and Sneppen, 2002) as it preserves the degree distribution in the network. However, a small-world network such as the brain is defined by a mean path length that is close to random characteristics (Milgram, 1967). As efficiency, which is a main measure used in this article, is computed based on the path length, the efficiency of a graph created by the Maslov–Sneppen algorithm will be very similar to that of the original graph. Thus, we created the random networks that preserves the group-average matrix edge weight distribution and not the degree distribution.

Statistical analysis and network analysis measures

Rich club and s-coreness

The rich-club coefficient is a global metric describing the extent to which nodes with high degrees are also connected between them. S-coreness is a weighted version of the coreness metric. For this metric, the maximal subgraph in which all nodes have a strength of k or higher is computed, and the coreness for each node is the highest k for which it can be found in the subgraph.

Both the rich-club coefficient Φw(k) and s-coreness were computed based on the group-average matrix as described by van den Heuvel and Sporns (2011). We generated both measures using the Brain Connectivity Toolbox for Python (Rubinov and Sporns, 2010).

As random graphs also display some rich-club properties, we estimated Φw(k) for each of the random normalization matrices mentioned above and set Φw(k)rand as the averaged rich-club coefficient over all of the random permutations. The normalized Φw(k) was defined as Φw(k)norm=Φw(k)Φw(k)rand. The p value for the difference between Φw(k) and Φw(k)rand for each k was the ratio of random matrices in which the rich-club coefficient was higher than that of the group-average matrix. The range of k presenting a significant difference between Φw(k) and Φw(k)rand is the node degrees characterizing the rich-club nodes

Mean shortest path and efficiency

To estimate the hemispheric connectivity, we calculated the network efficiency as described by Latora and Marchiori (2001). As the efficiency measure is normalized by the efficiency of an ideal graph, we first normalized each subject's connectivity matrix between 0 and 1 so that the ideal graph efficiency would be 1. The shortest path length (l) between each pair of nodes was computed using the Dijkstra (1959) algorithm, with the distance between each pair of connected nodes set as 1connection strength, that is, the more streamlines we found between the two areas i and j, the distance dij was smaller. The network efficiency was computed as the average of 1l for all node pairs. We repeated this for each hemisphere separately to include only association tracts and averaged the values for the overall hemispheric efficiency.

Commissural ratio

We estimated the commissural connectivity in two different ways. The percentage of commissural connections was calculated as the number of streamlines passing through the subject's corpus callosum. We created a 2D midsagittal corpus callosum mask for each subject. The masks were created using automatic region-of-interest selection based on a fiber orientation map and then manually checked and corrected them based on the T1 and fiber orientation distribution (FOD) maps where necessary. We then filtered the full brain tracts to only include tracts passing through the corpus callosum mask. To reduce the chance of including false-positive tracts, we also discarded tracts that recrossed the midsagittal plane outside the corpus callosum, as well as tracts that crossed between nonhomologous areas as most tracts in the corpus callosum are relatively symmetrical (Hofer and Frahm, 2006). The proportion of commissural connections was computed as the percent of streamlines that pass through this mask divided by the total number of streamlines for this subject.

As this proportion might be biased by tracking parameters, we used an additional estimate of the commissural ratio. This measure is the quotient of the square root of the midsagittal area of the corpus callosum (based on the corpus callosum mask) divided by the cubic root of the brain volume (based on the brain mask). This measure, although more commonly used (Manger et al., 2010), is affected by image resizing and also represents the corpus callosum area, without taking into account the fiber density in this area.

Hemispheric efficiency and commissural ratio correlation

Node and edge removal

To locate brain regions with a high effect on connectivity conservation, we removed each node (i) from the network and recalculated the hemispheric efficiency without that node (ei). We then compared the correlation between the commissural ratio and ei with the original correlation.

As strong connections are expected to have a greater effect on connectivity conservation, we also calculated the network efficiency value for networks with an increasing pruning threshold. That is, edges were removed by connectivity strength, with the weaker connections removed first. The hemispheric efficiency/commissural ratio correlation was computed after each pruning.

Heritability

Heritability measures quantify the extent to which people's genes account for differences in their traits. It is a statistical concept describing the level of variation in a specific trait that can be attributed to genetic variation. In this work, we estimated heritability using two measures. First, Falconer's formula (Falconer, 1960) defines heritability (h2) based on the difference between twin correlations in monozygotic and dizygotic twins for the examined trait [h2=2(rMZ−rDZ)]. Second, To assess the heritability of the connections between two traits, we used cross-twin-cross-trait correlations. In this method, the correlations between trait A in one twin and trait B in the second one are compared for monozygotic against dizygotic twins. Traits are considered genetically dependent if the cross-twin-cross-trait correlation is significantly higher for monozygotic than dizygotic twins. To avoid the effect of sex differences on the result, only same-sex dizygotic twins were used in the heritability analysis.

Canonical correlation analysis modeling for cognitive data and network measures

To examine the possible connections to cognitive abilities, we performed canonical correlation analysis (CCA). For this analysis, subjects with missing data in any of the measures used were discarded. As monozygotic twins were found to have a very similar network configuration, one twin of each monozygotic pair was also removed from the analysis.

An 819 × 5 (subjects × graph measures) matrix Mgraph was created. This matrix consisted of the subjects' efficiency, tractography-based and volume-based commissural ratio, residuals of the commissural ratio and hemispheric efficiency correlation, and the residuals' squared values.

A second matrix was created to represent the cognitive measures for each subject. Both corpus callosum size and functional and structural global network efficiency have been previously reported as associated with general intelligence (Luders et al., 2007; Kim et al., 2016; Pineda-Pardo et al., 2016). To examine this connection, we used cognitive measures associated with intelligence from the NIH Toolbox Cognition composite scores (Heaton et al., 2014). Each feature was standardized (z-score) across subjects.

To assess whether areas that have a strong effect on the connectivity conservation phenomenon also contribute to cognitive abilities, two similar, 819 × 17 (subjects × graph measures) matrices (Mstrong and Mweak) were created. These matrices included the s-coreness of the nodes with the strongest or weakest effect on the connectivity conservation, respectively. The canonical correlation between each of these matrices and the cognitive measures were compared (Table 1).

View this table:
  • View inline
  • View popup
Table 1.

All network, node, and cognitive measures used to create the matrices for the CCA

Canonical correlation was cross-validated by removing a random 25% of the subjects as test data. We extracted the canonical weights using the remaining 75% and then computed the canonical correlations for the test dataset, using these canonical weights. This process was repeated over 1000 permutations, choosing a random set of subjects for the test dataset each time. The r values reported here are the averaged out-of-sample correlations over all permutations.

Significance for each correlation was assessed by randomizing the subject's order in Mgraph relative to the cognitive matrix over 1000 permutations and rerunning the CCA for each. The p value was set as the ratio of canonical pairs that were more strongly correlated for the random matrices than the original ones.

To establish the effect of the conservation phenomenon on the canonical correlation, we examined which graph measures contributed most to the correlation. For that, we computed the canonical loadings of each measure. The canonical loadings are the correlation between the measure itself and the canonical components so that the closer they are to 1, the more effective the measure is on the overall correlation.

Results

Brain network characteristics

The network of the brain is expected to present a rich-club organization (van den Heuvel and Sporns, 2011; Grayson et al., 2014). As such, it should have a core of nodes with high degrees that are highly interconnected and constitute the main connection between the lower degree nodes. This brain organization has been examined and described previously, along with a characterization of specific nodes that form the rich-club network hubs (van den Heuvel and Sporns, 2011; Ball et al., 2014; Mišić et al., 2014).

Figure 1.
  • Download figure
  • Open in new tab
  • Download powerpoint
Figure 1.

Brain volume distribution and impact on connectivity and the rich-club coefficient over different degree thresholds. A, The rich-club coefficient for the HCP dataset subjects (dark blue), randomized networks (light blue), and the normalized value (red). The k values with p < 0.05 are marked in a gray background on the graph as the k values for the rich-club nodes. B, Brain volume seemed to have a small effect on overall connectivity in the brain, with longer characteristic path lengths for larger brains, although the difference in connectivity was minor. C, A histogram of the variations in brain size. The variability in brain volumes was large, with the largest brain nearly 1.8 times larger than the smallest one. Note that we use weighted networks (weighted by 1/# fibers) and thus the actual efficiency values are on an arbitrary scale.

The connectomes extracted for our subjects presented a rich-club organization. The rich-club coefficient Φw(k) is computed as a function of node degree cutoff (k). Our data show a rich-club organization, as expected, with an increase of normalized Φw(k) over increasing k cutoffs (Fig. 1A). The normalized Φw(k) drops at higher cutoffs, when enough nodes are excluded to lose the network structure. Nodes constituting the rich club include the anterior and middle cingulate cortex, superior frontal areas, medial occipital areas, the parahippocampal gyrus, postcentral and precentral gyri, and the precuneus (Fig. 1B).

Our recent study, focusing on the mammalian connectome, showed that although white matter volume appeared to be a function of brain volume, overall connectivity did not vary much across species (Assaf et al., 2020). This appeared to be consistent despite the large variance in brain volume and network structure, where the overall brain connectivity across mammals decreased by only 40% from the largest brain to the smallest brain, although brain volume decreased by more than four orders of magnitude.

Here, we examined the overall connectivity in the human brain and the dependency on brain volume. Overall brain network efficiency was estimated as the mean of the reciprocals of the distances (measured by the connection strength) between all node pairs. Pearson's correlation between whole-brain efficiency and brain volume for all the HCP subjects showed only a weak correlation (r=0.1,p=0.003). This was the case, despite the large variability in brain size between subjects, with a 1:1.8 ratio between the smallest and largest brain volume (Fig. 1C).

Figure 2.
  • Download figure
  • Open in new tab
  • Download powerpoint
Figure 2.

Hemispheric efficiency correlation to the commissural ratio. Correlation between the average of the efficiencies computed for each hemisphere and the commissural ratio was significant with r=0.−0.51 and p=0.1×10−62.

A correlation between hemispheric efficiency and commissural ratio

To examine the connection between commissural connection strength and hemispheric connectivity, we estimated Pearson's correlation between the commissural ratio and hemispheric efficiency. A strong and significant correlation was found between the two measures (r=−0.51, p=0.1×10−62; Fig. 2). When computing the hemispheric efficiency based on the randomized brain matrices, the hemispheric efficiency and commissural ratio correlation was insignificant (r=0.003, p=0.94)

The parcellation used in this analysis is an atlas-based anatomic parcellation. To negate possible biases of the specific parcellation (de Reus and van den Heuvel, 2013), we also created an anatomy-free parcellation using k-means clustering of all tract endpoints in each hemisphere for each subject separately. This created three unique parcellations for each subject, with 150, 300, or 500 areas per hemisphere that are not based on known anatomic or functional priors. Connectivity matrices were created based on these parcellations, and efficiency was computed for each hemisphere with similar results.

To ensure the correlation is not a result of our protocol, we extracted new connectomes for each subject using different numbers of tracts and received similar results. Normalizing the connection strength for each edge between 0 and 1 and setting the distance between nodes as −log(#fibers) also had little effect on the results.

As connection strength was set by the number of streamlines found connecting every two areas, weaker connections, that is, connections consisting of fewer streamlines, have a higher probability to be false positives. To negate possible biases from false tracts, we tested the effect of pruning by testing various pruning thresholds and found that connectivity conservation is robust across most thresholds.

Node-based analysis

Different areas have a different effect on overall brain connectivity and on the balance between interhemispheric and intrahemispheric connections. To examine whether specific regions in the brain contribute more to the overall trade-off, we reran the analysis on a leave-one-out basis, removing each node from the network and recalculating the hemispheric efficiency to observe the effect of removal of that node on the trade-off between efficiency and the commissural ratio.

Removing the hippocampus, the superior motor areas, and the ACC from the network resulted in a lower correlation, suggesting that they are strong contributors to the effect. When removing other nodes such as occipital and temporal areas around the insula, we observed a higher negative correlation, suggesting they are nodes with connections that reduce the effect. These regions consist of a large part of the rich club and is mostly characterized by a high centrality in the network measured by the s-coreness (van den Heuvel and Sporns, 2011), which is a measure of the centrality of the node in the network. A significant correlation (r=0.39, p=0.0003) was found between the removal effect of the node on the correlation and s-coreness (Fig. 3). Furthermore, removing nodes by the s-corness threshold shows a large drop in correlation when removing central nodes and little to no effect when removing the weaker ones.

Figure 3.
  • Download figure
  • Open in new tab
  • Download powerpoint
Figure 3.

Contribution of nodes to the overall trade-off. Nodes with the strongest effect on the correlation between hemispheric efficiency and commissural ratio as found in the leave-one-out analysis. A, Nodes in each hemisphere colored by the Δr between the correlations with and without that node. Nodes with a high centrality were more likely to affect the overall balance. B, All brain regions in the parcellation by descending s-coreness, colored by the Δr between the correlations with and without that node. C, There is a correlation between Δr and the s-coreness of the dropped node.

Heritability

Some elements in the connectome structure have a genetic source, whereas others are affected by the environment during brain development. As the connectivity conservation phenomenon seems to be a fundamental principle of the network architecture, and is inherent across species, we expected this trade-off to have an apparent heritable source. Participants were divided into monozygotic and dizygotic twins, pairs of siblings, and third siblings of twin pairs. Participants in the last group were randomly divided into unrelated pairs. We used Tukey's HSD to examine the differences in network efficiency and in the commissural ratio between these related pairs. The difference in the commissural ratio in monozygotic twins was significantly lower than that of all other pairs (MZ/DZ, p = 0.004; MZ/not twin, p = 0.04; MZ/not siblings, p = 0.001). However, the difference in efficiency between monozygotic twin pairs was only significantly lower than that of the nontwin and unrelated pairs (p = 0.04 and p = 0.001, respectively; Fig. 4).

Figure 4.
  • Download figure
  • Open in new tab
  • Download powerpoint
Figure 4.

The difference in hemispheric efficiency and the commissural ratio between sibling pairs. A, B, The difference in (A) hemispheric efficiency and (B) commissural ratio between sibling pairs. The x-axis signifies the mean absolute difference between siblings across groups. Corpus callosum size and diffusion measures are hypothesized to be affected by heritability factors.

Both measures were highly heritable, with Falconer's h2 for the commissural ratio showing 99% heritability and for efficiency showing 88% heritability, demonstrating the strong effect of heredity on the network structure of the connectome.

Although this might point to a stronger environmental effect on the hemispheric efficiency than the commissural ratio, the connection between the two measures, that is, the basis of the connectivity conservation phenomenon presented in this work, seems to be genetically influenced. The cross-twin-cross-trait correlation between the commissural ratio and efficiency was significant for monozygotic twins (r=−0.51,p=0.01) and not for dizygotic twins (r=−0.2,p=0.28). This result points to a genetic connection between the two factors.

Canonical correlation analysis for cognitive abilities

To examine the cognitive relevance of the connectivity conservation phenomenon, we ran a CCA analysis. This method examines the cross-covariance between two sets matrices (structured subjects × variables, with different variables in each matrix) by finding pairs of linear combinations for each set of variables in each matrix that have the largest correlation between pairs and with the minimum correlation between linear combinations of the same set. Here we examined the correlation between a cognitive measures matrix (Mcog) and matrices describing sets of connectivity measures.

The first canonical pair for Mcog X Mgraph (all overall network measures), showed a significant correlation (r = 0.13, p = 0.013). Each feature in each matrix was assigned a weight based on the canonical loading of the feature, that is, the correlation between the original variable and the canonical variate of that set. These weights represent a measure of the contribution of this variable to the canonical correlation. An examination of the canonical loadings of the correlation showed strong contributions by measures associated with crystalized intelligence, that is, the picture vocabulary and oral reading scores as well as the crystallized cognition composite score (Weintraub et al., 2013; Fig. 5A).

Figure 5.
  • Download figure
  • Open in new tab
  • Download powerpoint
Figure 5.

Canonical correlation between connectivity measures and crystallized intelligence. Color and bar heights represent the r value of the correlation between the cognitive measure and the canonical variate. A, Canonical loadings for each cognitive measure used for the initial CCA. Measures associated with crystalized cognition are marked in red. B, Correlation between the canonical variates for the overall measures and the cognitive ones. C, Canonical loadings for the nodes with the strongest contribution to the connectivity conservation phenomenon. Nodes are presented by order of contribution and colored by the Δr between the correlations with and without that node.

We further examined the canonical correlation between connectivity measures (Mgraph) and crystalized intelligence (Mcryst−cog). We found a significant canonical correlation for the first canonical pair (r = 0.15, p = 0.003; Fig. 5B). When repeating the analysis with the fluid intelligence measures, the canonical correlation was both lower and not significant (r = 0.07, p = 0.073).

When examining the canonical loadings for Mgraph × Mcryst−cog, the tractography-based commissural ratio had the highest loadings (−0.77), followed by efficiencies (0.31). The residual measures also had notable loadings (−0.18 for squared residuals).

Last, we examined the canonical correlation for Mcryst−cog and a matrix consisting of the s-coreness of nodes with a high contribution to the connectivity conservation phenomenon (Mstrong; nodes contribution was assessed by the difference in the hemispheric connectivity-commissural ratio when this node is removed). The canonical correlation between Mcryst−cog and Mstrong was significant (r = 0.11, p = 0.012), whereas the correlation for the least contributing nodes (Mweak) was insignificant (r = 0.02, p = 0.06). High canonical loadings were observed for bilateral motor areas, frontal inferior areas, and the lingual gyrus as well as the right middle occipital area and the left calcarine sulcus (Fig. 5C).

Discussion

In this study, we explored the existence and impact and of a connectivity conservation phenomenon in a large cohort of humans as seen before across mammals. We further study the mechanisms at the base of the connectivity conservation across brains.

Network analysis is an emerging theme in neuroimaging in the past few years (Bassett and Sporns, 2017). Recent works focus on specialized network structures or specific populations. As such, many works have explored modular structures of brain networks, local subnetworks, and differences in community structure (Valencia et al., 2009; Meunier et al., 2010; Sporns and Betzel, 2016; Puxeddu et al., 2020), while others have examined the dynamics of the network (Bassett et al., 2011; Beaty et al., 2016; Sporns, 2018). We present a cohort-based network analysis research, spanning a large variety of samples and exploring the overall architecture of the human connectome. This large-scale overview provides a new outlook on the network structure of the brain.

We show a trade-off between the association and commissural fibers, where brains with higher intrahemispheric connectivity have fewer connections between hemispheres, explaining the connectivity consistency seen across subjects and species. Our results pinpoint the biological origins of this phenomenon, with central rich-club nodes at the basis. Furthermore, we explore possible hereditary sources of the phenomenon and the effect on cognitive abilities. Our findings may shed light on the factors that guide brain network architectures and the effect of these factors.

Previous works show that white matter volume is dependent on brain size (Zhang and Sejnowski, 2000). However, our results show that brain efficiency is independent of brain volume and thus on white matter volume. Furthermore, our results show a correlation between the efficiency of the hemispheric network and the proportion of commissural fibers in the brain, indicating a conserved balance between these two factors. These findings strengthen the hypothesis that brain networks are organized in a way that preserves overall consistent connectivity.

Examining this phenomenon within species, and especially within the human brain, allows us to more easily examine the contribution of the network structure to connectivity conservation. As the variability in the human brain structure is smaller than it is among different species, this comparison supports the results because it shows that the vast differences in brain structure and network architecture across species are not the underlying cause of this trade-off. The human brain is also well researched and well mapped, allowing us to look into the effect of different structures on this phenomenon on a comparable parcellation.

When examining the brain regions that may have a stronger contribution to the connectivity conservation, our results show a strong effect of the hippocampus, the superior motor and occipital areas, and the areas surrounding the insula. Most of these areas are considered as network hubs or rich-club nodes (van den Heuvel and Sporns, 2011, 2013), that is, highly interconnected areas that form the connections between other nodes. As such, these areas are expected to be central in connectivity mechanisms.

We also explored the heritability of commissural and hemispheric connections using twin and sibling data. Earlier results observed a high heritability of the corpus callosum (Johnson et al., 1994; Woldehawariat et al., 2014). Many studies have also examined the heritability of diffusion measures and streamline count in the brain (Thompson et al., 2013; Budisavljevic et al., 2016; Yeh et al., 2016) as well as the functional connectome (Miranda-Dominguez et al., 2018). However, the heritability of the structural connectome and the overall connectivity measures does not seem to be widely investigated.

In the developing brain, individual differences in both functional and structural connectivity are reflected in the brain function. These differences may be inherent in the network structure, pointing to a genetic source, or change over time, and reflect plasticity due to each individual's specific experiences. Here, although the factors leading a specific brain to be on each end of the connectivity balance scale are unknown, a basis for the organizing mechanism for the whole network is seen. By showing heritability in both commissural ratio and hemispheric efficiency, as well as a strong cross-twin cross-trait correlation, our work points to a hereditary source for connectivity conservation. This, along with findings across different mammalian species, shows that this phenomenon is not environment or even species specific.

The brain network structure, consisting of a network of hubs strongly connected together, creates a modular structure that is at the base of network efficiency (van den Heuvel and Sporns, 2011). Functional studies have shown that performance in crystalized cognitive tasks is associated with the modular state of the functional network (Bertolero et al., 2015; Shine et al., 2016). These works suggest that the modular architecture of the network, and thus the efficiency, has an effect on the individual's cognitive abilities (Barbey, 2018).

The connectome architecture displays individually unique characteristics across subjects (Finn et al., 2015). These features have been seen to predict cognitive functions (Cai et al., 2021), and some were hypothesized to be affected by individual experiences (Marques et al., 2015). Our results also highlight a connection between the connectivity architecture and intersubject variability in cognitive abilities. Our analysis points to a connection between the components of the connectivity conservation phenomenon and crystallized intelligence. We also observe a weaker connection between the residuals and crystallized intelligence measures, suggesting that incompliance with the interhemispheric and intrahemispheric balance might affect cognitive abilities. It is difficult to infer the direction of connection between specific variables in the CCA. However, correlation direction suggests that lower hemispheric efficiency is associated with a higher crystallized cognition score. Corpus callosum size and diffusion measures have been previously found to be associated with intelligence and other cognitive functions (Hines et al., 1992; Clarke and Zaidel, 1994; Luders et al., 2007). Furthermore, brain efficiency was associated with intelligence (Kim et al., 2016; Ma et al., 2017), and deficits in higher cognitive abilities have been associated with deficiencies in brain network structure (Brown and Paul, 2000; Rudie et al., 2012) and efficiency (Du et al., 2019). These findings may underline our results, as the cognitive variance between subjects with different values of hemispheric connectivity is apparent. Our findings may shed light on the mechanism behind the intersubject cognitive variability.

The connectome extraction process has many parameters to be considered, and the resulting connectome is somewhat sensitive to those parameters. To avoid bias based on brain volume or resolution, all parameters were normalized to the volume of each brain. The number of nodes and the variability of these nodes in size may also create a bias in the connectivity measures (Cammoun et al., 2012) as well as using an anatomic parcellation. To avoid a bias toward larger areas, similar size areas were used. For the anatomic parcellation bias, the same analysis was run with a non-prior-based atlas, with different numbers of areas. All parcellations led to similar results, rejecting the possibility of the parcellation affecting the outcome. Another limitation of using tractography to map the corpus callosum tracts is the relatively large number of false positives in the tract-creating process. Manual validation and filtering of the corpus callosum tracts can easily locate many of those false streamlines but is more susceptible to variance because of subjective decisions. Here, as a compromise between the need for manual validation and conformity among samples, we mapped the common false tracts manually and created a set of objective rules to guide automatic filtering.

Our findings suggest that a trade-off between hemispheric connectivity and interhemispheric connections is the basis of the consistent connectivity of the brain network over different architectures and brain volumes. These results provide a unique opportunity to examine the network characteristics underlying this phenomenon and map the biological origin. Furthermore, our results examine the cognitive implications of this phenomenon.

As a connectivity phenomenon that spans across subjects, connectivity conservation should be examined in many scopes. First, as brain connectivity changes drastically over these years, it is important to examine changes in overall brain connectivity over time and the predictability of the adults brain network structure from that of a young age. Next, as functional networks are highly correlated to structural networks, exploring the phenomenon and the implication in the functional networks is essential, especially as the dynamics of the functional network and the ability of the network to change from a more connected to a more modular state (Barbey, 2018) may in part explain the intersubject differences in cognitive abilities. Furthermore, although this study has examined a large cohort of subjects, the difference in connectivity conservation in smaller populations, such as different age groups or subjects with pathologies, may show a different effect in the network structure. Last, our findings regarding cognitive measures connected to the phenomenon could lead to the exploration of this connectivity conservation mechanism in specialized networks or the changes in the connection structure induced by learning.

Although further studies should continue to explore this phenomenon and look into the cognitive implications, this study provides a deeper insight into the basis of the network structure of the brain and characterizes the architecture of the hemispheric network in the brain.

Footnotes

  • This work was supported by Israel Science Foundation Grant 1303/20 (to Y.A.), National Science Foundation–Israel Binational Science Foundation Collaborative Research in Computational Neuroscience Grant 2018711 (to Y.A. and Y.Y.), and British Council Israel Research and Academic Exchange Grant 43BX (to Y.A.). Data were provided by the Human Connectome Project, Washington University and University of Minnesota Consortium (1U54MH091657), funded by the 16 National Institutes of Health that support the Blueprint for Neuroscience Research, and by the McDonnell Center for Systems Neuroscience at Washington University. We thank Prof. Olaf Sporns from Indiana University for discussions.

  • The authors declare no competing financial interests.

  • Correspondence should be addressed to Ronnie Krupnik at ronniek{at}mail.tau.ac.il

SfN exclusive license.

References

  1. ↵
    1. Achard S,
    2. Bullmore E
    (2007) Efficiency and cost of economical brain functional networks. PLOS Comput Biol 3:e17. e17. doi:10.1371/journal.pcbi.0030017 pmid:17274684
    OpenUrlCrossRefPubMed
  2. ↵
    1. Achard S,
    2. Salvador R,
    3. Whitcher B,
    4. Suckling J,
    5. Bullmore E
    (2006) A resilient, low-frequency, small-world human brain functional network with highly connected association cortical hubs. J Neurosci 26:63–72. doi:10.1523/JNEUROSCI.3874-05.2006 pmid:16399673
    OpenUrlAbstract/FREE Full Text
  3. ↵
    1. Assaf Y,
    2. Bouznach A,
    3. Zomet O,
    4. Marom A,
    5. Yovel Y
    (2020) Conservation of brain connectivity and wiring across the mammalian class. Nat Neurosci 23:805–808. doi:10.1038/s41593-020-0641-7 pmid:32514137
    OpenUrlCrossRefPubMed
  4. ↵
    1. Ball G,
    2. Aljabar P,
    3. Zebari S,
    4. Tusor N,
    5. Arichi T,
    6. Merchant N,
    7. Robinson EC,
    8. Ogundipe E,
    9. Rueckert D,
    10. Edwards AD,
    11. Counsell SJ
    (2014) Rich-club organization of the newborn human brain. Proc Natl Acad Sci U S A 111:7456–7461. doi:10.1073/pnas.1324118111 pmid:24799693
    OpenUrlAbstract/FREE Full Text
  5. ↵
    1. Barbey AK
    (2018) Network neuroscience theory of human intelligence. Trends Cogn Sci 22:8–20. doi:10.1016/j.tics.2017.10.001 pmid:29167088
    OpenUrlCrossRefPubMed
  6. ↵
    1. Bassett DS,
    2. Bullmore ET
    (2009) Human brain networks in health and disease. Curr Opin Neurol 22:340–347. doi:10.1097/WCO.0b013e32832d93dd pmid:19494774
    OpenUrlCrossRefPubMed
  7. ↵
    1. Bassett DS,
    2. Sporns O
    (2017) Network neuroscience. Nat Neurosci 20:353–364. doi:10.1038/nn.4502 pmid:28230844
    OpenUrlCrossRefPubMed
  8. ↵
    1. Bassett DS,
    2. Wymbs NF,
    3. Porter MA,
    4. Mucha PJ,
    5. Carlson JM,
    6. Grafton ST
    (2011) Dynamic reconfiguration of human brain networks during learning. Proc Natl Acad Sci U S A 108:7641–7646. doi:10.1073/pnas.1018985108 pmid:21502525
    OpenUrlAbstract/FREE Full Text
  9. ↵
    1. Beaty RE,
    2. Benedek M,
    3. Silvia PJ,
    4. Schacter DL
    (2016) Creative cognition and brain network dynamics. Trends Cogn Sci 20:87–95. doi:10.1016/j.tics.2015.10.004 pmid:26553223
    OpenUrlCrossRefPubMed
  10. ↵
    1. Bertolero MA,
    2. Yeo BTT,
    3. D'Esposito M
    (2015) The modular and integrative functional architecture of the human brain. Proc Natl Acad Sci U S A 112:E6798–E6807. doi:10.1073/pnas.1510619112 pmid:26598686
    OpenUrlAbstract/FREE Full Text
  11. ↵
    1. Brown WS,
    2. Paul LK
    (2000) Cognitive and psychosocial deficits in agenesis of the corpus callosum with normal intelligence. Cogn Neuropsychiatry 5:135–157. doi:10.1080/135468000395781
    OpenUrlCrossRef
  12. ↵
    1. Budisavljevic S,
    2. Kawadler JM,
    3. Dell F,
    4. Rijsdijk FV,
    5. Kane F,
    6. Picchioni M,
    7. McGuire P,
    8. Toulopoulou T,
    9. Georgiades A,
    10. Kalidindi S,
    11. Kravariti E,
    12. Murray RM,
    13. Murphy DG,
    14. Craig MC,
    15. Catani M
    (2016) Heritability of the limbic networks. Soc Cogn Affect Neurosci 11:746–757. doi:10.1093/scan/nsv156 pmid:26714573
    OpenUrlCrossRefPubMed
  13. ↵
    1. Bullmore E,
    2. Sporns O
    (2012) The economy of brain network organization. Nat Rev Neurosci 13:336–349. doi:10.1038/nrn3214 pmid:22498897
    OpenUrlCrossRefPubMed
  14. ↵
    1. Bullmore ET,
    2. Bassett DS
    (2011) Brain graphs: graphical models of the human brain connectome. Annu Rev Clin Psychol 7:113–140. doi:10.1146/annurev-clinpsy-040510-143934 pmid:21128784
    OpenUrlCrossRefPubMed
  15. ↵
    1. Cai B,
    2. Zhang G,
    3. Zhang A,
    4. Xiao L,
    5. Hu W,
    6. Stephen JM,
    7. Wilson TW,
    8. Calhoun VD,
    9. Wang Y-P
    (2021) Functional connectome fingerprinting: identifying individuals and predicting cognitive functions via autoencoder. Hum Brain Mapp 42:2691–2705. doi:10.1002/hbm.25394 pmid:33835637
    OpenUrlCrossRefPubMed
  16. ↵
    1. Cammoun L,
    2. Gigandet X,
    3. Meskaldji D,
    4. Thiran JP,
    5. Sporns O,
    6. Do KQ,
    7. Maeder P,
    8. Meuli R,
    9. Hagmann P
    (2012) Mapping the human connectome at multiple scales with diffusion spectrum MRI. J Neurosci Methods 203:386–397. doi:10.1016/j.jneumeth.2011.09.031 pmid:22001222
    OpenUrlCrossRefPubMed
  17. ↵
    1. Clarke JM,
    2. Zaidel E
    (1994) Anatomical-behavioral relationships: corpus callosum morphometry and hemispheric specialization. Behav Brain Res 64:185–202. doi:10.1016/0166-4328(94)90131-7 pmid:7840886
    OpenUrlCrossRefPubMed
  18. ↵
    1. de Reus MA,
    2. van den Heuvel MP
    (2013) The parcellation-based connectome: limitations and extensions. Neuroimage 80:397–404. doi:10.1016/j.neuroimage.2013.03.053 pmid:23558097
    OpenUrlCrossRefPubMed
  19. ↵
    1. Dhollander T,
    2. Raffelt D,
    3. Connelly A
    (2016) Unsupervised 3-tissue response function estimation from single-shell or multi-shell diffusion MR data without a co-registered T1 image. Paper presented ISMRM Workshop on Breaking the Barriers of Diffusion MRI, Lisbon, September.
  20. ↵
    1. Dijkstra E
    (1959) A note on two problems in connexion with graphs. Numer Math 1:269–271. doi:10.1007/BF01386390
    OpenUrlCrossRef
  21. ↵
    1. Du J,
    2. Wang Y,
    3. Zhi N,
    4. Geng J,
    5. Cao W,
    6. Yu L,
    7. Mi J,
    8. Zhou Y,
    9. Xu Q,
    10. Wen W,
    11. Sachdev P
    (2019) Structural brain network measures are superior to vascular burden scores in predicting early cognitive impairment in post stroke patients with small vessel disease. Neuroimage Clin 22:101712. doi:10.1016/j.nicl.2019.101712
    OpenUrlCrossRef
  22. ↵
    1. Falconer D
    (1960) Introduction to Quantitative Genetics. Ronald Press, New York.
  23. ↵
    1. Finn ES,
    2. Shen X,
    3. Scheinost D,
    4. Rosenberg MD,
    5. Huang J,
    6. Chun MM,
    7. Papademetris X,
    8. Constable RT
    (2015) Functional connectome fingerprinting: identifying individuals using patterns of brain connectivity. Nat Neurosci 18:1664–1671. doi:10.1038/nn.4135 pmid:26457551
    OpenUrlCrossRefPubMed
  24. ↵
    1. Fischer FU,
    2. Wolf D,
    3. Scheurich A,
    4. Fellgiebel A
    (2014) Association of structural global brain network properties with intelligence in normal aging. PLoS One 9:e86258. doi:10.1371/journal.pone.0086258 pmid:24465994
    OpenUrlCrossRefPubMed
  25. ↵
    1. Gershon RC,
    2. Cella D,
    3. Fox NA,
    4. Havlik RJ,
    5. Hendrie HC,
    6. Wagster MV
    (2010) Assessment of neurological and behavioural function: the NIH Toolbox. Lancet Neurol 9:138–139. doi:10.1016/S1474-4422(09)70335-7 pmid:20129161
    OpenUrlCrossRefPubMed
  26. ↵
    1. Glasser MF,
    2. Sotiropoulos SN,
    3. Wilson JA,
    4. Coalson TS,
    5. Fischl B,
    6. Andersson JL,
    7. Xu J,
    8. Jbabdi S,
    9. Webster M,
    10. Polimeni JR,
    11. Van Essen DC,
    12. Jenkinson M
    . (2013) The minimal preprocessing pipelines for the Human Connectome Project. Neuroimage 80:105–124. doi:10.1016/j.neuroimage.2013.04.127 pmid:23668970
    OpenUrlCrossRefPubMed
  27. ↵
    1. Grayson DS,
    2. Ray S,
    3. Carpenter S,
    4. Iyer S,
    5. Dias TGC,
    6. Stevens C,
    7. Nigg JT,
    8. Fair DA
    (2014) Structural and functional rich club organization of the brain in children and adults. PLoS One 9:e88297. doi:10.1371/journal.pone.0088297 pmid:24505468
    OpenUrlCrossRefPubMed
  28. ↵
    1. He Y,
    2. Chen ZJ,
    3. Evans AC
    (2007) Small-world anatomical networks in the human brain revealed by cortical thickness from MRI. Cereb Cortex 17:2407–2419. doi:10.1093/cercor/bhl149 pmid:17204824
    OpenUrlCrossRefPubMed
  29. ↵
    1. Heaton RK,
    2. Akshoomoff N,
    3. Tulsky D,
    4. Mungas D,
    5. Weintraub S,
    6. Dikmen S,
    7. Beaumont J,
    8. Casaletto KB,
    9. Conway K,
    10. Slotkin J,
    11. Gershon R
    (2014) Reliability and validity of composite scores from the NIH toolbox cognition battery in adults. J Int Neuropsychol Soc 20:588–598. doi:10.1017/S1355617714000241 pmid:24960398
    OpenUrlCrossRefPubMed
  30. ↵
    1. Hines M,
    2. Chiu L,
    3. McAdams LA,
    4. Bentler PM,
    5. Lipcamon J
    (1992) Cognition and the corpus callosum: verbal fluency, visuospatial ability, and language lateralization related to midsagittal surface areas of callosal subregions. Behav Neurosci 106:3–14. doi:10.1037/0735-7044.106.1.3 pmid:1554435
    OpenUrlCrossRefPubMed
  31. ↵
    1. Hofer S,
    2. Frahm J
    (2006) Topography of the human corpus callosum revisited—comprehensive fiber tractography using diffusion tensor magnetic resonance imaging. Neuroimage 32:989–994. doi:10.1016/j.neuroimage.2006.05.044 pmid:16854598
    OpenUrlCrossRefPubMed
  32. ↵
    1. Jenkinson M,
    2. Bannister P,
    3. Brady M,
    4. Smith S
    (2002) Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage 17:825–841. doi:10.1016/s1053-8119(02)91132-8 pmid:12377157
    OpenUrlCrossRefPubMed
  33. ↵
    1. Jeurissen B,
    2. Tournier JD,
    3. Dhollander T,
    4. Connelly A,
    5. Sijbers J
    (2014) Multi-tissue constrained spherical deconvolution for improved analysis of multi-shell diffusion MRI data. Neuroimage 103:411–426. doi:10.1016/j.neuroimage.2014.07.061 pmid:25109526
    OpenUrlCrossRefPubMed
  34. ↵
    1. Johnson SC,
    2. Farnworth T,
    3. Pinkston JB,
    4. Bigler ED,
    5. Blatter DD
    (1994) Corpus callosum surface area across the human adult life span: effect of age and gender. Brain Res Bull 35:373–377. doi:10.1016/0361-9230(94)90116-3 pmid:7850489
    OpenUrlCrossRefPubMed
  35. ↵
    1. Kim D-J,
    2. Davis EP,
    3. Sandman CA,
    4. Sporns O,
    5. O'Donnell BF,
    6. Buss C,
    7. Hetrick WP
    (2016) Children's intellectual ability is associated with structural network integrity. Neuroimage 124:550–556. doi:10.1016/j.neuroimage.2015.09.012 pmid:26385010
    OpenUrlCrossRefPubMed
  36. ↵
    1. Latora V,
    2. Marchiori M
    (2001) Efficient behavior of small-world networks. Phys Rev Lett 87:198701–198704. doi:10.1103/PhysRevLett.87.198701 pmid:11690461
    OpenUrlCrossRefPubMed
  37. ↵
    1. Luders E,
    2. Narr KL,
    3. Bilder RM,
    4. Thompson PM,
    5. Szeszko PR,
    6. Hamilton L,
    7. Toga AW
    (2007) Positive correlations between corpus callosum thickness and intelligence. Neuroimage 37:1457–1464. doi:10.1016/j.neuroimage.2007.06.028 pmid:17689267
    OpenUrlCrossRefPubMed
  38. ↵
    1. Ma J,
    2. Kang HJ,
    3. Kim JY,
    4. Jeong HS,
    5. Im JJ,
    6. Namgung E,
    7. Kim MJ,
    8. Lee S,
    9. Kim TD,
    10. Oh JK,
    11. Chung YA,
    12. Lyoo IK,
    13. Lim SM,
    14. Yoon S
    (2017) Network attributes underlying intellectual giftedness in the developing brain. Sci Rep 7:11321. doi:10.1038/s41598-017-11593-3 pmid:28900176
    OpenUrlCrossRefPubMed
  39. ↵
    1. Manger PR,
    2. Hemingway J,
    3. Haagensen M,
    4. Gilissen E
    (2010) Cross-sectional area of the elephant corpus callosum: comparison to other eutherian mammals. Neuroscience 167:815–824. doi:10.1016/j.neuroscience.2010.02.066 pmid:20206234
    OpenUrlCrossRefPubMed
  40. ↵
    1. Marques P,
    2. Soares JM,
    3. Magalhães R,
    4. Santos NC,
    5. Sousa N
    (2015) The bounds of education in the human brain connectome. Sci Rep 6:12812–12818. doi:10.1038/srep12812 pmid:26245669
    OpenUrlCrossRefPubMed
  41. ↵
    1. Maslov S,
    2. Sneppen K
    (2002) Specificity and stability in topology of protein networks. Science 296:910–913. doi:10.1126/science.1065103 pmid:11988575
    OpenUrlAbstract/FREE Full Text
  42. ↵
    1. Meunier D,
    2. Lambiotte R,
    3. Bullmore ET
    (2010) Modular and hierarchically modular organization of brain networks. Front Neurosci 4:200. doi:10.3389/fnins.2010.00200 pmid:21151783
    OpenUrlCrossRefPubMed
  43. ↵
    1. Milgram S
    (1967) The small world problem. Psychol Today 2:60–67.
    OpenUrl
  44. ↵
    1. Miranda-Dominguez O,
    2. Feczko E,
    3. Grayson DS,
    4. Walum H,
    5. Nigg JT,
    6. Fair DA
    (2018) Heritability of the human connectome: a connectotyping study. Netw Neurosci 2:175–199. doi:10.1162/netn_a_00029 pmid:30215032
    OpenUrlCrossRefPubMed
  45. ↵
    1. Mišić B,
    2. Sporns O,
    3. McIntosh AR
    (2014) Communication efficiency and congestion of signal traffic in large-scale brain networks. PLoS Comput Biol 10:e1003427. doi:10.1371/journal.pcbi.1003427 pmid:24415931
    OpenUrlCrossRefPubMed
  46. ↵
    1. Owen JP,
    2. Li YO,
    3. Yang FG,
    4. Shetty C,
    5. Bukshpun P,
    6. Vora S,
    7. Wakahiro M,
    8. Hinkley LBN,
    9. Nagarajan SS,
    10. Sherr EH,
    11. Mukherjee P
    (2013a) Resting-state networks and the functional connectome of the human brain in agenesis of the corpus callosum. Brain Connect 3:547–562. doi:10.1089/brain.2013.0175 pmid:24063289
    OpenUrlCrossRefPubMed
  47. ↵
    1. Owen JP,
    2. Li YO,
    3. Ziv E,
    4. Strominger Z,
    5. Gold J,
    6. Bukhpun P,
    7. Wakahiro M,
    8. Friedman EJ,
    9. Sherr EH,
    10. Mukherjee P
    (2013b) The structural connectome of the human brain in agenesis of the corpus callosum. Neuroimage 70:340–355. doi:10.1016/j.neuroimage.2012.12.031 pmid:23268782
    OpenUrlCrossRefPubMed
  48. ↵
    1. Patenaude B,
    2. Smith SM,
    3. Kennedy DN,
    4. Jenkinson M
    (2011) A Bayesian model of shape and appearance for subcortical brain segmentation. Neuroimage 56:907–922. doi:10.1016/j.neuroimage.2011.02.046 pmid:21352927
    OpenUrlCrossRefPubMed
  49. ↵
    1. Pineda-Pardo JA,
    2. Martínez K,
    3. Román FJ,
    4. Colom R
    (2016) Structural efficiency within a parieto-frontal network and cognitive differences. Intelligence 54:105–116. doi:10.1016/j.intell.2015.12.002
    OpenUrlCrossRef
  50. ↵
    1. Puxeddu MG,
    2. Faskowitz J,
    3. Betzel RF,
    4. Petti M,
    5. Astolfi L,
    6. Sporns O
    (2020) The modular organization of brain cortical connectivity across the human lifespan. Neuroimage 218:116974. doi:10.1016/j.neuroimage.2020.116974 pmid:32450249
    OpenUrlCrossRefPubMed
  51. ↵
    1. Rubinov M,
    2. Sporns O
    (2010) Complex network measures of brain connectivity: Uses and interpretations. NeuroImage 52:1059–1069.
    OpenUrlCrossRefPubMed
  52. ↵
    1. Rudie JD,
    2. Brown JA,
    3. Beck-Pancer D,
    4. Hernandez LM,
    5. Dennis EL,
    6. Thompson PM,
    7. Bookheimer SY,
    8. Dapretto M
    (2012) Altered functional and structural brain network organization in autism. NeuroImage Clin 2:79–94. doi:10.1016/j.nicl.2012.11.006 pmid:24179761
    OpenUrlCrossRefPubMed
  53. ↵
    1. Shine JM,
    2. Bissett PG,
    3. Bell PT,
    4. Koyejo O,
    5. Balsters JH,
    6. Gorgolewski KJ,
    7. Moodie CA,
    8. Poldrack RA
    (2016) The dynamics of functional brain networks: integrated network states during cognitive task performance. Neuron 92:544–554. doi:10.1016/j.neuron.2016.09.018 pmid:27693256
    OpenUrlCrossRefPubMed
  54. ↵
    1. Smith RE,
    2. Tournier JD,
    3. Calamante F,
    4. Connelly A
    (2012) Anatomically-constrained tractography: improved diffusion MRI streamlines tractography through effective use of anatomical information. Neuroimage 62:1924–1938. doi:10.1016/j.neuroimage.2012.06.005 pmid:22705374
    OpenUrlCrossRefPubMed
  55. ↵
    1. Smith RE,
    2. Tournier JD,
    3. Calamante F,
    4. Connelly A
    (2013) SIFT: spherical-deconvolution informed filtering of tractograms. Neuroimage 67:298–312. doi:10.1016/j.neuroimage.2012.11.049 pmid:23238430
    OpenUrlCrossRefPubMed
  56. ↵
    1. Smith SM,
    2. Jenkinson M,
    3. Woolrich MW,
    4. Beckmann CF,
    5. Behrens TEJ,
    6. Johansen-Berg H,
    7. Bannister PR,
    8. De Luca M,
    9. Drobnjak I,
    10. Flitney DE,
    11. Niazy RK,
    12. Saunders J,
    13. Vickers J,
    14. Zhang Y,
    15. De Stefano N,
    16. Brady JM,
    17. Matthews PM
    (2004) Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 23 Suppl 1:S208–S219. doi:10.1016/j.neuroimage.2004.07.051 pmid:15501092
    OpenUrlCrossRefPubMed
  57. ↵
    1. Sotiropoulos SN,
    2. Zalesky A
    (2019) Building connectomes using diffusion MRI: why, how and but. NMR Biomed 32:e3752. doi:10.1002/nbm.3752 pmid:28654718
    OpenUrlCrossRefPubMed
  58. ↵
    1. Sporns O
    (2018) Graph theory methods: applications in brain networks. Dialogues Clin Neurosci 20:111–120. pmid:30250388
    OpenUrlPubMed
  59. ↵
    1. Sporns O,
    2. Betzel RF
    (2016) Modular Brain Networks: a set of elements (nodes) and the pairwise interactions (edges) among those elements. Annu Rev Psychol 67:613–653. doi:10.1146/annurev-psych-122414-033634 pmid:26393868
    OpenUrlCrossRefPubMed
  60. ↵
    1. Sporns O,
    2. Chialvo DR,
    3. Kaiser M,
    4. Hilgetag CC
    (2004) Organization, development and function of complex brain networks. Trends Cogn Sci 8:418–425. doi:10.1016/j.tics.2004.07.008 pmid:15350243
    OpenUrlCrossRefPubMed
  61. ↵
    1. Stam CJ
    (2014) Modern network science of neurological disorders. Nat Rev Neurosci 15:683–695. doi:10.1038/nrn3801 pmid:25186238
    OpenUrlCrossRefPubMed
  62. ↵
    1. Stam C,
    2. Jones B,
    3. Nolte G,
    4. Breakspear M,
    5. Scheltens P
    (2007) Small-world networks and functional connectivity in Alzheimer's Disease. Cereb Cortex 17:92–99. doi:10.1093/cercor/bhj127 pmid:16452642
    OpenUrlCrossRefPubMed
  63. ↵
    1. Thompson PM,
    2. Ge T,
    3. Glahn DC,
    4. Jahanshad N,
    5. Nichols TE
    (2013) Genetics of the connectome. Neuroimage 80:475–488. doi:10.1016/j.neuroimage.2013.05.013 pmid:23707675
    OpenUrlCrossRefPubMed
  64. ↵
    1. Tournier JD,
    2. Calamante F,
    3. Gadian DG,
    4. Connelly A
    (2004) Direct estimation of the fiber orientation density function from diffusion-weighted MRI data using spherical deconvolution. Neuroimage 23:1176–1185. doi:10.1016/j.neuroimage.2004.07.037 pmid:15528117
    OpenUrlCrossRefPubMed
  65. ↵
    1. Tournier JD,
    2. Smith R,
    3. Raffelt D,
    4. Tabbara R,
    5. Dhollander T,
    6. Pietsch M,
    7. Christiaens D,
    8. Jeurissen B,
    9. Yeh CH,
    10. Connelly A
    (2019) MRtrix3: a fast, flexible and open software framework for medical image processing and visualisation. Neuroimage 202:116137. doi:10.1016/j.neuroimage.2019.116137 pmid:31473352
    OpenUrlCrossRefPubMed
  66. ↵
    1. Tzourio-Mazoyer N,
    2. Landeau B,
    3. Papathanassiou D,
    4. Crivello F,
    5. Etard O,
    6. Delcroix N,
    7. Mazoyer B,
    8. Joliot M
    (2002) Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage 15:273–289. doi:10.1006/nimg.2001.0978 pmid:11771995
    OpenUrlCrossRefPubMed
  67. ↵
    1. Valencia M,
    2. Pastor MA,
    3. Fernández-Seara MA,
    4. Artieda J,
    5. Martinerie J,
    6. Chavez M
    (2009) Complex modular structure of large-scale brain networks. Chaos 19:023119. doi:10.1063/1.3129783 pmid:19566254
    OpenUrlCrossRefPubMed
  68. ↵
    1. van den Heuvel MP,
    2. Sporns O
    (2011) Rich-club organization of the human connectome. J Neurosci 31:15775–15786. doi:10.1523/JNEUROSCI.3539-11.2011 pmid:22049421
    OpenUrlAbstract/FREE Full Text
  69. ↵
    1. van den Heuvel MP,
    2. Sporns O
    (2013) Network hubs in the human brain. Trends Cogn Sci 17:683–696. doi:10.1016/j.tics.2013.09.012 pmid:24231140
    OpenUrlCrossRefPubMed
  70. ↵
    1. van den Heuvel MP,
    2. Stam CJ,
    3. Boersma M,
    4. Hulshoff Pol HE
    (2008) Small-world and scale-free organization of voxel-based resting-state functional connectivity in the human brain. Neuroimage 43:528–539. doi:10.1016/j.neuroimage.2008.08.010 pmid:18786642
    OpenUrlCrossRefPubMed
  71. ↵
    1. Van Essen DC,
    2. Ugurbil K,
    3. Auerbach E,
    4. Barch D,
    5. Behrens TEJ,
    6. Bucholz R,
    7. Chang A,
    8. Chen L,
    9. Corbetta M,
    10. Curtiss SW,
    11. Della Penna S,
    12. Feinberg D,
    13. Glasser MF,
    14. Harel N,
    15. Heath AC,
    16. Larson-Prior L,
    17. Marcus D,
    18. Michalareas G,
    19. Moeller S,
    20. Oostenveld R, et al
    . (2012) The Human Connectome Project: a data acquisition perspective. Neuroimage 62:2222–2231. doi:10.1016/j.neuroimage.2012.02.018 pmid:22366334
    OpenUrlCrossRefPubMed
  72. ↵
    1. Vlooswijk MCG,
    2. Vaessen MJ,
    3. Jansen JFA,
    4. De Krom M,
    5. Majoie HJM,
    6. Hofman PAM,
    7. Aldenkamp AP,
    8. Backes WH
    (2011) Loss of network efficiency associated with cognitive decline in chronic epilepsy. Neurology 77:938–944. doi:10.1212/WNL.0b013e31822cfc2f pmid:21832213
    OpenUrlCrossRefPubMed
  73. ↵
    1. Weintraub S,
    2. Dikmen SS,
    3. Heaton RK,
    4. Tulsky DS,
    5. Zelazo PD,
    6. Bauer PJ,
    7. Carlozzi NE,
    8. Slotkin J,
    9. Blitz D,
    10. Wallner-Allen K,
    11. Fox NA,
    12. Beaumont JL,
    13. Mungas D,
    14. Nowinski CJ,
    15. Richler J,
    16. Deocampo JA,
    17. Anderson JE,
    18. Manly JJ,
    19. Borosh B,
    20. Havlik R, et al.
    (2013) Cognition assessment using the NIH Toolbox. Neurology 80:S54–S64. doi:10.1212/WNL.0b013e3182872ded pmid:23479546
    OpenUrlAbstract/FREE Full Text
  74. ↵
    1. Wen W,
    2. Zhu W,
    3. He Y,
    4. Kochan NA,
    5. Reppermund S,
    6. Slavin MJ,
    7. Brodaty H,
    8. Crawford J,
    9. Xia A,
    10. Sachdev P
    (2011) Discrete neuroanatomical networks are associated with specific cognitive abilities in old age. J Neurosci 31:1204–1212. doi:10.1523/JNEUROSCI.4085-10.2011 pmid:21273405
    OpenUrlAbstract/FREE Full Text
  75. ↵
    1. Wiseman SJ,
    2. Booth T,
    3. Ritchie SJ,
    4. Cox SR,
    5. Muñoz Maniega S,
    6. Valdés Hernández M. d C,
    7. Dickie DA,
    8. Royle NA,
    9. Starr JM,
    10. Deary IJ,
    11. Wardlaw JM,
    12. Bastin ME
    (2018) Cognitive abilities, brain white matter hyperintensity volume, and structural network connectivity in older age. Hum Brain Mapp 39:622–632. doi:10.1002/hbm.23857 pmid:29139161
    OpenUrlCrossRefPubMed
  76. ↵
    1. Woldehawariat G,
    2. Martinez PE,
    3. Hauser P,
    4. Hoover DM,
    5. Drevets WWC,
    6. McMahon FJ
    (2014) Corpus callosum size is highly heritable in humans, and may reflect distinct genetic influences on ventral and rostral regions. PLoS One 9:e99980. doi:10.1371/journal.pone.0099980
    OpenUrlCrossRef
  77. ↵
    1. Yeh CH,
    2. Smith RE,
    3. Dhollander T,
    4. Calamante F,
    5. Connelly A
    (2019) Connectomes from streamlines tractography: assigning streamlines to brain parcellations is not trivial but highly consequential. Neuroimage 199:160–171. doi:10.1016/j.neuroimage.2019.05.005 pmid:31082471
    OpenUrlCrossRefPubMed
  78. ↵
    1. Yeh F-C,
    2. Vettel JM,
    3. Singh A,
    4. Poczos B,
    5. Grafton ST,
    6. Erickson KI,
    7. Tseng W-YI,
    8. Verstynen TD
    (2016) Quantifying differences and similarities in whole-brain white matter architecture using local connectome fingerprints. PLOS Comput Biol 12:e1005203. doi:10.1371/journal.pcbi.1005203 pmid:27846212
    OpenUrlCrossRefPubMed
  79. ↵
    1. Zhang K,
    2. Sejnowski TJ
    (2000) A universal scaling law between gray matter and white matter of cerebral cortex. Proc Natl Acad Sci U S A 97:5621–5626. doi:10.1073/pnas.090504197 pmid:10792049
    OpenUrlAbstract/FREE Full Text
  80. ↵
    1. Zhang Y,
    2. Brady M,
    3. Smith S
    (2001) Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm. IEEE Trans Med Imaging 20:45–57. doi:10.1109/42.906424 pmid:11293691
    OpenUrlCrossRefPubMed
  81. ↵
    1. Zuo XN,
    2. He Y,
    3. Betzel RF,
    4. Colcombe S,
    5. Sporns O,
    6. Milham MP
    (2017) Human connectomics across the life span. Trends Cogn Sci 21:32–45. doi:10.1016/j.tics.2016.10.005 pmid:27865786
    OpenUrlCrossRefPubMed
Back to top

In this issue

The Journal of Neuroscience: 41 (40)
Journal of Neuroscience
Vol. 41, Issue 40
6 Oct 2021
  • Table of Contents
  • Table of Contents (PDF)
  • About the Cover
  • Index by author
  • Ed Board (PDF)
Email

Thank you for sharing this Journal of Neuroscience article.

NOTE: We request your email address only to inform the recipient that it was you who recommended this article, and that it is not junk mail. We do not retain these email addresses.

Enter multiple addresses on separate lines or separate them with commas.
Inner Hemispheric and Interhemispheric Connectivity Balance in the Human Brain
(Your Name) has forwarded a page to you from Journal of Neuroscience
(Your Name) thought you would be interested in this article in Journal of Neuroscience.
CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.
Print
View Full Page PDF
Citation Tools
Inner Hemispheric and Interhemispheric Connectivity Balance in the Human Brain
Ronnie Krupnik, Yossi Yovel, Yaniv Assaf
Journal of Neuroscience 6 October 2021, 41 (40) 8351-8361; DOI: 10.1523/JNEUROSCI.1074-21.2021

Citation Manager Formats

  • BibTeX
  • Bookends
  • EasyBib
  • EndNote (tagged)
  • EndNote 8 (xml)
  • Medlars
  • Mendeley
  • Papers
  • RefWorks Tagged
  • Ref Manager
  • RIS
  • Zotero
Respond to this article
Request Permissions
Share
Inner Hemispheric and Interhemispheric Connectivity Balance in the Human Brain
Ronnie Krupnik, Yossi Yovel, Yaniv Assaf
Journal of Neuroscience 6 October 2021, 41 (40) 8351-8361; DOI: 10.1523/JNEUROSCI.1074-21.2021
Twitter logo Facebook logo Mendeley logo
  • Tweet Widget
  • Facebook Like
  • Google Plus One

Jump to section

  • Article
    • Abstract
    • Introduction
    • Materials and Methods
    • Results
    • Discussion
    • Footnotes
    • References
  • Figures & Data
  • Info & Metrics
  • eLetters
  • PDF

Keywords

  • connectivity
  • imaging
  • MRI
  • network
  • tractography

Responses to this article

Respond to this article

Jump to comment:

No eLetters have been published for this article.

Related Articles

Cited By...

More in this TOC Section

Research Articles

  • Increased neuronal expression of the early endosomal adaptor APPL1 replicates Alzheimer’s Disease-related endosomal and synaptic dysfunction with cholinergic neurodegeneration.
  • Presynaptic mu opioid receptors suppress the functional connectivity of ventral tegmental area dopaminergic neurons with aversion-related brain regions
  • Change of spiny neuron structure in the basal ganglia song circuit and its regulation by miR-9 during song development
Show more Research Articles

Systems/Circuits

  • Presynaptic mu opioid receptors suppress the functional connectivity of ventral tegmental area dopaminergic neurons with aversion-related brain regions
  • V2b neurons act via multiple targets to produce in phase inhibition during locomotion
  • The Neurobiology of Cognitive Fatigue and Its Influence on Effort-Based Choice
Show more Systems/Circuits
  • Home
  • Alerts
  • Follow SFN on BlueSky
  • Visit Society for Neuroscience on Facebook
  • Follow Society for Neuroscience on Twitter
  • Follow Society for Neuroscience on LinkedIn
  • Visit Society for Neuroscience on Youtube
  • Follow our RSS feeds

Content

  • Early Release
  • Current Issue
  • Issue Archive
  • Collections

Information

  • For Authors
  • For Advertisers
  • For the Media
  • For Subscribers

About

  • About the Journal
  • Editorial Board
  • Privacy Notice
  • Contact
  • Accessibility
(JNeurosci logo)
(SfN logo)

Copyright © 2025 by the Society for Neuroscience.
JNeurosci Online ISSN: 1529-2401

The ideas and opinions expressed in JNeurosci do not necessarily reflect those of SfN or the JNeurosci Editorial Board. Publication of an advertisement or other product mention in JNeurosci should not be construed as an endorsement of the manufacturer’s claims. SfN does not assume any responsibility for any injury and/or damage to persons or property arising from or related to any use of any material contained in JNeurosci.