Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Sep 1.
Published in final edited form as: J Neurosci Methods. 2024 Jun 22;409:110205. doi: 10.1016/j.jneumeth.2024.110205

Global brain connectivity: test-retest stability and association with biological and neurocognitive variables

Gianpaolo Del Mauro 1, Yiran Li 1, Ze Wang 1
PMCID: PMC11286348  NIHMSID: NIHMS2005796  PMID: 38914376

Abstract

Background

Global brain connectivity (GBC) enables measuring brain regions’ functional connectivity strength at rest by computing the average correlation between each brain voxel’s time-series and that of all other voxels.

New method

We used resting-state fMRI (rs-fMRI) data of young adult participants from the Human Connectome Project (HCP) dataset to explore the test-retest stability of GBC, the brain regions with higher or lower GBC, as well as the associations of this measure with age, sex, and fluid intelligence. GBC was computed by considering separately the positive and negative correlation coefficients (positive GBC and negative GBC).

Results

Test-retest stability was higher for positive compared to negative GBC. Areas with higher GBC were located in the default mode network, insula, and visual areas, while regions with lower GBC were in subcortical regions, temporal cortex, and cerebellum. Higher age was related to global reduction of positive GBC. Males displayed higher positive GBC in the whole brain. Fluid intelligence was associated to increased positive GBC in fronto-parietal, occipital and temporal regions.

Comparison with Existing Method

Compared to previous works, this study adopted a larger sample size and tested GBC stability using data from different rs-fMRI sessions. Moreover, these associations were examined by testing positive and negative GBC separately.

Conclusions

Lower stability for negative compared to positive GBC suggests that negative correlations may reflect less stable couplings between brain regions. Our findings indicate a greater importance of positive compared to negative GBC for the associations of functional connectivity strength with biological and neurocognitive variables.

Keywords: Global Brain Connectivity, resting-state fMRI, Fluid intelligence, Age, Sex

1. Introduction

In the last decades, resting-state functional magnetic resonance imaging (rs-fMRI) has been widely adopted to investigate spontaneous brain activity without stimuli or tasks. Among the techniques developed to analyze rs-fMRI data, functional connectivity (FC) has gained a prominent role enabling to measure the synchronization of the spontaneous neuronal activity between anatomically distinct brain regions (Lv et al., 2018) with high test-retest stabilities (Braun et al., 2012; Chen et al., 2015; Li et al., 2012). Numerous studies have investigated FC using different techniques to detect distinct types of information regarding the functional coupling of brain regions, such as FC density (FCD) (Tomasi and Volkow, 2010), seed-based FC (Lv et al., 2018), independent component analysis (ICA) (Biswal et al., 1995), and graph analysis (Bullmore and Sporns, 2009).

Instead of focusing on preselected seed regions or networks, each voxel can be compared to the rest of the brain to get its global connectivity and find the brain’s most connected regions. Such a global brain connectivity (GBC) analysis (Cole et al., 2010), similar to the weighted degree centrality in the context of graph theory (Zuo et al., 2012) or the mean FC within the task-invoked brain regions (Wang et al., 2012), computes the linear correlation of each voxel in the brain with all other voxels and summarizes that to obtain a single value representing the average FC strength (the terms “GBC”, “FC strength” and “correlation strength” will be used here interchangeably). Previous studies have demonstrated the potential neurocognitive correlates of the regional GBC patterns but were based on moderate sample size.

The purpose of this study is to use large high quality and densely sampled rsfMRI data from the Human Connectome Project (HCP) – Young Adult dataset (Van Essen et al., 2013) to characterize GBC through the following analyses: we examined the populational GBC distributions for the positive GBC (p-GBC) and negative GBC (n-GBC); we then explored the biological association of the GBC maps with age and sex; lastly, we investigated the neurocognitive correlates of GBC by examining its association with nonverbal fluid intelligence (nv-F), which includes aspects of general intelligence independent of knowledge and culture. The HCP rsfMRI data were acquired using the state-of-art multi-band imaging sequence. We have recently analyzed them and reported physiologically and biologically meaningful results even after controlling a few confounds including motions (Camargo et al., 2024; Del Mauro and Wang, 2024, 2023; Del Mauro et al., 2024; Lin et al., 2022; Wang, 2021).

Prior findings indicate fronto-parietal and default mode network (DMN) regions as the brain areas with the highest GBC (Cole et al., 2010), and that higher age is associated with reduced within- and between-network correlation strength in many sensory and cognitive brain networks (Siman-Tov et al., 2017; Varangis et al., 2019). GBC does not differ between males and females (Salvador et al., 2020), despite previous findings suggesting sex differences in brain FC (Satterthwaite et al., 2015). Meanwhile, using both functional and structural MRI techniques and in accordance with the parieto-frontal integration theory (P-FIT) of intelligence (Jung and Haier, 2007), fluid intelligence has been repeatedly associated with fronto-parietal regions overlapping with resting-state networks such as the dorsal attention network (DAN) and executive control network (ECN) (Santarnecchi et al., 2017a, 2017b; Tadayon et al., 2020). These findings have been partially confirmed by other studies reporting an association between fluid intelligence and GBC in the left prefrontal cortex (Cole et al., 2015, 2012). Based on these results, we hypothesize that: i) fronto-parietal and DMN regions will show higher FC strength compared to other brain areas (Cole et al., 2010); ii) age will be negatively associated with GBC (Siman-Tov et al., 2017; Varangis et al., 2019); iii) males and females will not differ in terms of FC strength (Salvador et al., 2020); iv) nv-F will be associated with GBC of fronto-parietal regions (Jung and Haier, 2007; Santarnecchi et al., 2017b). Moreover, the acquisition of multiple rs-fMRI sessions in the HCP (Van Essen et al., 2013) will allow to test the consistency of the associations between GBC and age, sex, and nv-F as well as the test-retest stability of GBC estimated from different rs-fMRI sessions.

2. Materials and methods

2.1. Informed consent

Data acquisition and sharing have been approved by the HCP parent IRB. Written informed consent forms have been obtained from all subjects before any experiments. This study reanalyzed the HCP data and data Use Terms have been signed and approved by the WU-Minn HCP Consortium.

2.2. Data

Fully pre-processed rs-fMRI, demographic, and behavioral data were downloaded from the Human Connectome Project (HCP) – Young Adult dataset (Van Essen et al., 2013). Participants had undergone two rs-fMRI sessions (Rest1 and Rest2, respectively) in two separate days, each including two runs. Each run was acquired using the same multi-band sequence (Moeller et al., 2010) and included 1200 timepoints. To compensate for image distortions induced by the long scan time, readout direction was from left to right for the first run of each session (Rest1 LR and Rest2 LR), and from right to left for the second run of each session (Rest1 RL and Rest2 RL). All data were acquired using the same scanning protocol, including: repetition time (TR) = 720ms, echo time (TE) = 33.1ms, resolution = 2x2x2mm3. The HCP pre-processing pipeline included the following steps: correction for spatial distortions, head motion, and B0 distortions, registration to T1w structural image, normalization to MNI space, global intensity normalization, masking out of non-brain voxels, and removal of confounds and non-neuronal artifacts (e.g., physiological signal) through high-pass temporal filtering and independent component analysis (ICA) (a detailed description of the pre-processing pipeline can be found in Glasser et al., 2013 and Smith et al., 2013). Only data of participants who completed four rs-fMRI were download from the HCP, leading to a sample of 1018 participants (mean age = 28.73 ± 3.7 years, range = 22-37; males/females = 472/546).

2.3. Global Brain Connectivity (GBC)

A graphical summary of the pipeline followed to calculate GBC is reported in Fig. 1. Note that correction for head motion artifacts was already applied during HCP pre-processing pipeline using ICA-based approach (see above). During GBC calculation, potential residual motion effects will be global and will be removed through the normalization procedure. As for signal dropouts, these occur in the gradient weighted EPI images in areas with large susceptibility gradients such as the boundary of the brain and the air-tissue conjunction area. Signal dropouts, however, are less problematic in the HCP data because of the short readout due to acceleration and good field shimming. We haven’t observed severe signal dropout in the data included in this work. GBC was computed independently for all rs-fMRI runs as the average connectivity of each voxel (Fig. 1a). First, the Pearson’s correlation coefficient of each brain voxel with all other voxels (excluding self-correlations) was calculated to obtain a connectivity matrix. To reduce computation time, white matter and cerebrospinal fluid were masked out and only the upper triangle of the matrix was calculated (Fig. 1b). For each within-brain voxel, two values were extracted: the average of the positive correlation coefficients (i.e., p-GBC), and the average of the absolute values of negative correlation coefficients (i.e., n-GBC) (Fig. 1c). For each of the two categories, values of all voxels were re-scaled by multiplying by 1000 to form the final GBC image.

Fig. 1.

Fig. 1

Open in a new tab

Pipeline followed to calculate global brain connectivity (GBC). For each participant: a) GBC was independently computed for four rs-fMRI runs; b) Pearson’s correlation coefficient of each brain voxel with all other voxels (excluding self-correlations) was calculated to obtain a connectivity matrix (CM). To reduce computation time, only the upper triangle of the matrix was calculated; c) for each within-brain voxel, two values were extracted: the average of the positive correlation coefficients (i.e., p-GBC), and the average of the absolute values of negative correlation coefficients (i.e., n-GBC). Values of all voxels were re-scaled by multiplying by 1000 to form the final GBC image; d) GBC maps obtained from the runs of the first session (Rest1 LR and Rest1 RL) and the GBC maps obtained from the runs of the second session (Rest2 LR and Rest2 RL) were averaged and smoothed with a Gaussian filter with full-width-at-half-maximum (FWHM) = 4 mm

The GBC maps obtained from the runs of the first session (Rest1 LR and Rest1 RL) and the GBC maps obtained from the runs of the second session (Rest2 LR and Rest2 RL) were averaged and smoothed with a Gaussian filter with full-width-at-half-maximum (FWHM) = 4 mm using the Nilearn library running on Python 3 (Abraham et al., 2014), resulting in four maps for each participant: two from Rest1 and two from Rest2, each including a positive and a negative map (Fig. 1d).

2.4. Nonverbal fluid intelligence (nv-F)

Fluid intelligence is a construct that measures higher-order relational functionality for solving problems in novel situations independent of acquired knowledge (Cattell, 1963). In the HCP, nonverbal fluid intelligence (nv-F) was measured of the day of Rest1 by adopting a shortened version of the commonly used Raven’s Progressive Matrices (Raven, 2000), namely the Penn Matrix Test (PMT) (Barch et al., 2013; Bilker et al., 2012; Moore et al., 2015), that includes 24 items. In each test item, a visual geometric design with a missing piece is presented, and the subject must select out of five possible choices the missing element that correctly completes the pattern. Finding the right missing piece relies on abstract reasoning and logic thinking rather than learned skills or knowledge (e.g., language, reading, or writing). In the following statistical analyses, the number of correct answers at the PMT will be used as a measure of nv-F.

2.5. Statistical analyses

2.5.1. Test-retest stability

The two-way random average score intraclass correlation coefficient (ICC) was used to assess the test-retest stability of the GBC between Rest1 and Rest2. For each of the two GBC maps, the ICC was calculated between each voxel from the Rest1 map and the corresponding voxel of the Rest2 map. Two ICC maps, one for each GBC map, were obtained and thresholded at ICC ≥ 0.3.

2.5.2. Regional differences in GBC

All analyses were performed using the Nilearn library (Abraham et al., 2014), which enables to compute the desired t-contrasts and produces Z-statistical brain maps as outputs. Using a built-in function, Z-statistical brain maps can be thresholded by calculating the Z-threshold corresponding to the desired p-threshold. In this study, all results were thresholded at p < 0.05 corrected for multiple comparisons using the false-discovery rate (FDR). For this analysis, GBC maps collected at the Rest1 session (averaged across the first LR and the RL scans) and the Rest2 session (averaged across the second LR and the RL scans) were converted to Z-scores (mean = 0, standard deviation =1) before smoothing (see Materials and Methods), and then a FWHM filter of 4 mm was applied. One-sample t-tests were run to investigate brain regions with FC strength significantly higher or lower than the group mean. Analyses were performed for each GBC map and for Rest1 and Rest2 separately.

2.5.3. Biological and cognitive associations of GBC

General linear models (GLM) were run to test the effects of age, sex, and nv-F on positive and negative GBC maps averaged from Rest1 and Rest2. First, a GLM including the GBC maps as dependent variable and age and sex as regressors was estimated to test the effects of age and sex. The variable sex was codified as follows: females = 1 and males = −1. Then, nv-F was included into the model to test its effect on GBC. Four participants who did not complete the Penn Matrix Test were excluded from this analysis (N = 1014). Z-statistic brain images resulting from the analyses were thresholded at p-FDR< 0.05 and clusters size > 200 voxels.

3. Results

3.1. Test-retest stability

p-GBC showed high ICC in the entire brain (mean ICC = 0.71) (Fig. 2a). n-GBC showed relatively lower reliability across the whole brain (mean ICC = 0.48) (Fig. 2b). For both positive and negative correlation strength, lower ICC values were displayed in the basal ganglia, orbitofrontal cortex, medial temporal lobes, and cerebellum.

Fig. 2.

Fig. 2

Open in a new tab

Intraclass correlation coefficient (ICC) was used to assess the test-retest stability of the positive (a) and negative (b) global brain connectivity (GBC). For each of the two GBC maps, the ICC was calculated between each voxel from the Rest1 map and the corresponding voxel of the Rest2 map. Darker colors indicate higher ICC. Only ICCs ≥ 0.3 are shown in the figure

3.2. Regional differences in GBC

Rest1 and Rest2 GBC maps showed identical results, therefore only Rest1 findings are presented in Fig. 3. Overall, dorsolateral prefrontal cortex (PFC), middle temporal gyri as well as parietal and occipital lobes had higher p-GBC (Rest1 Z-threshold = 1.97, min beta = 0.02; Rest2 Z-threshold = 1.97, min beta = 0.04), while inferior frontal gyri, ventromedial PFC, insula, temporal lobes, subcortical regions, and cerebellum had lower p-GBC (Rest1 Z-threshold = 1.97, min beta = −0.02; Rest2 Z-threshold = 1.97, min beta = −0.02). The dorsomedial PFC and the anterior/middle cingulate cortex showed a less defined pattern, with sulcal and gyral regions displaying, respectively, lower and higher positive correlation strength. For the n-GBC, regions with higher FC strength, corresponding to stronger negative connections, were mainly identified in the dorsolateral, dorsomedial and ventromedial PFC, inferior and middle temporal gyri, hippocampus, inferior parietal lobules (IPL), precuneus/posterior cingulate cortex (PCu/PCC), occipital poles, and superior-caudal regions of the cerebellum (Rest1 Z-threshold = 1.99, min beta = 0.01; Rest2 Z-threshold = 1.99, min beta = 0.01). On the other hand, inferior frontal gyri, anterior/middle cingulate and somatomotor cortex, superior parietal lobules (SPL), superior temporal gyrus, occipital and medial temporal lobes, subcortical regions, and inferior-rostral regions of the cerebellum displayed lower negative FC strength (Rest1 Z-threshold = 1.99, min beta = −0.01; Rest2 Z-threshold = 1.99, min beta = −0.02) (Fig. 3b).

Fig. 3.

Fig. 3

Open in a new tab

One-sample t-tests performed on a) positive and b) negative global brain connectivity (GBC) maps from Rest1 (averaged across the first LR and the RL scans). Analyses on GBC maps calculated from Rest2 session yielded identical results (not shown). Color bar is based on Z-values. Significant results were identified by thresholding Z-values with a false-discovery rate (FDR) corrected p-value < 0.05. Only above-threshold results are showed. Upper and lower limits of the color bar are set as the highest Z-value across all analyses. For each voxel, higher Z-values (darker colors) indicate greater GBC (i.e., higher average of correlation coefficients), while lower Z-values (lighter colors) indicate weaker GBC (i.e., lower average of correlation coefficients). c) Binary map showing regions with higher (red) or lower (blue) GBC independently from the correlation sign

The intersection between panels a) and b) of Fig. 2 was computed to obtain a binary map showing regions with a statistically higher or lower GBC independently from the correlation sign (Fig. 3c). Regions showing higher GBC included dorsolateral and dorsomedial PFC, PCu, insula, inferior and superior parietal lobules, occipital lobes, and middle temporal gyri, while regions showing lower GBC included inferior frontal gyri, somatomotor cortex, anterior and middle cingulate cortex, lateral and medial temporal lobes, subcortical regions, and cerebellum (Fig. 3c).

3.3. Biological and cognitive associations of GBC

Age increase was related to a significant global reduction of the p-GBC in Rest1 (Z-threshold = 2.044, min beta = −0.13) (Fig. 3a) and Rest2 (Z-threshold = 1.971, min beta = −0.1) (Fig. 3b). No significant effect was observed for the n-GBC (p-FDR > 0.05).

Males compared to females showed a globally higher p-GBC in Rest1 (p-GBC: Z-threshold = 1.969, min beta = −0.03) and Rest2 (p-GBC: Z-threshold = 1.99, min beta = −0.03) (Fig. 4a, c). In addition, females displayed higher n-GBC in the left middle frontal gyrus, right insula, IPL, and cingulate cortex (Rest1 n-GBC: Z-threshold = 3.274, min beta = 0.83; Rest2 n-GBC: Z-threshold = 3.036, min beta = 0.44), whereas males had greater n-GBC in the cerebellum, right orbitofrontal cortex, and left hippocampus (Rest1 n-GBC: Z-threshold = 3.274, min beta = −0.48; Rest2 n-GBC: Z-threshold = 3.036, min beta = −0.48) (Fig. 4b, d).

Fig. 4.

Fig. 4

Open in a new tab

Effect of Age on the positive global brain connectivity (GBC) computed from a) Rest1 (averaged across the first LR and the RL scans) and b) Rest2 (averaged across the second LR and the RL scans) sessions. Age was globally related to reduced GBC. Color bar is based on Z-values. Darker colors indicate higher Z-values. Significant results were identified by thresholding Z-values with a false-discovery rate (FDR) corrected p-value < 0.05 and using a cluster threshold > 200 voxels. Lower limit of the color bar is set as the highest Z-value across all analyses

As for nv-F, in Rest1 (Z-threshold = 2.34, min beta = 0.13) this measure was related to a broad increase of the p-GBC in cortical and subcortical regions, including lateral and medial PFC, SPL, IPL, PCu/PCC, temporal and occipital lobes, caudate nucleus, putamen, and cerebellum (Fig. 5a). In the Rest2 session (Z-threshold = 3.16, min beta = 0.24), the effect of nv-F was limited to increased p-GBC only in the lateral PFC, IPL, right temporal cortex, and cerebellum (Fig. 5b). No significant effect was observed for the n-GBC.

Fig. 5.

Fig. 5

Open in a new tab

Sex effect on the positive and negative global brain connectivity (GBC) computed from (a, b) Rest1 (averaged across the first LR and the RL scans) and (c, d) Rest2 (averaged across the second LR and the RL scans) sessions. Color bar is based on Z-values. Darker colors indicate higher Z-values. Significant results were identified by thresholding Z-values with a false-discovery rate (FDR) corrected p-value < 0.05 and using a cluster threshold > 200 voxels. Upper and lower limits of the color bar are set as the highest Z-value across all analyses. Hot and cold colors indicate, respectively, higher GBC in females and higher GBC in males (p-FDR > 0.05)

4. Discussion

In this study, we examined the properties of GBC and its association with biological and neurocognitive variables, including age, sex, and nv-F using the large HCP-Youth data. Results will be discussed in detail in the following paragraphs.

4.1. Test-retest stability

Test-retest stability of FC strength was explored between GBC maps estimated from different rs-fMRI sessions. P-GBC displayed high ICC across the brain, while n-GBC was much less stable, showing low-to-moderate ICC values. This finding confirms previous studies indicating a higher variability and a lower reproducibility of negative correlations between brain regions (Chang and Glover, 2010; Shehzad et al., 2009), and suggests that negative correlations might reflect state-dependent couplings, while positive ones are more likely to represent anatomically constrained couplings between brain regions (Gopinath et al., 2015). Additionally, lower ICCs were generally observed in the basal ganglia, orbitofrontal cortex, medial temporal lobes, and cerebellum, consistently with results of other works examining reliability and reproducibility of FC (Noble et al., 2017; Shah et al., 2016).

4.2. Regional differences in GBC

Stronger positive correlation strength was broadly observed in frontal, parietal, and occipital lobes. Additionally, dorsolateral and medial PFC, inferior and middle temporal gyri, hippocampus, IPL, PCu/PCC, and superior-caudal regions of the cerebellum displayed stronger n-GBC. Interestingly, areas with high p-GBC did not clearly resemble any of the well-known resting-state brain networks, while most of the regions with high n-GBC are considered core components of the DMN (Raichle, 2015). This finding might reflect the repeatedly reported anti correlation between the DMN and other task-positive networks (Parente and Colosimo, 2020), which has also been related to higher cognitive functioning (Reineberg et al., 2018).

When examining GBC without considering the correlation sign, regions with lower GBC were mainly localized in subcortical regions, temporal cortex, and cerebellum. This result is overall consistent with previous studies on FCD where no or few connectivity hubs were reported in these regions (Tomasi et al., 2016). Moreover, it is worth noting that regions with weaker FC strength roughly corresponded to those with lower test-retest stability (see above). Prior findings suggest that stability of FC may depend upon different factors, including the strength of the correlation (Noble et al., 2019; Shehzad et al., 2009), and this might explain why regions with low GBC also tended to show lower stability. Intriguingly, we also found that many areas showing high GBC overlap with regions of DMN, including dorsolateral and dorsomedial PFC, precuneus, inferior parietal lobules, and lateral temporal cortex. Regions of the DMN showed the highest GBC in previous studies (Cole et al., 2010), and many of these areas have been indicated as functional connectivity hubs of brain networks characterized by high metabolic consumption (Bagarinao et al., 2020; Liang et al., 2013; Tomasi et al., 2016, 2013). Brain hubs are generally characterized by their high degree of connectivity to other regions (van den Heuvel and Sporns, 2013), and play a crucial role in information transfer and between-network communication within the brain (Sporns et al., 2007). Our results suggest that these regions also show stronger FC strength, highlighting their pivotal contribution to information flow within the brain.

4.3. Age

Higher age was related to reduced p-GBC across the whole brain and this finding represents the first to be reported within a sample of young adult participants (22-37 years). In previous studies, FC strength within most brain networks was found to decrease from late adolescence to young adulthood (16-26 years) (Marek et al., 2015) and from young adulthood (21-40 years) to middle age (41-60 years) (Siman-Tov et al., 2017). In a more recent study with a different age-based group categorization, FC strength of brain networks was not significantly different between young adults (20-34 years) and middle-aged adults (35-49 years) (Varangis et al., 2019). However, in the same study, using participants spanning from young adults to elders, age was negatively correlated with positive correlation weights in many sensory and cognitive networks. Overall, results reported in this study suggest that decrease in FC strength is a process that starts during late adolescence and continues in younger middle-aged adults. Significantly, negative correlation strength was not related to age. This might depend on the fact that negative correlations between brain regions are generally weaker and less likely to reflect anatomically constrained spontaneous fluctuations of brain activity (Chang and Glover, 2010; Gopinath et al., 2015).

4.4. Sex

Males displayed higher p-GBC across the whole brain, and higher n-GBC in the cerebellum, right orbitofrontal cortex, and hippocampus. Females showed only higher n-GBC in the left middle frontal gyrus, right insula, IPL, and cingulate cortex. This is the first time such a large difference between males and females has been reported using FC. Prior studies reported bidirectional differences with, e.g., males showing higher between-network and females higher within-network FC (Satterthwaite et al., 2015). Moreover, in a previous work females showed higher local FCD (Tomasi and Volkow, 2012), while no sex differences have been reported when examining FC strength (Salvador et al., 2020; Satterthwaite et al., 2015). Considering the inconsistency with previous results, potentially related to the larger sample size used in this work, these data do not enable at the moment to hypothesize potential biological mechanisms or sociocultural factors underlying these differences, which should be replicated in future works.

4.5. nv-F

Association between GBC and nv-F was identified only for p-GBC. As for the age effect, we suggest that a lacking relationship between n-GBC and nv-F might be related to the less stability and strength of negative compared to positive correlations. In Rest1, higher nv-F was related to GBC in the lateral and medial PFC, SPL, IPL, PCu/PCC, temporal and occipital lobes, caudate nucleus, putamen, and cerebellum. Overall, these findings are in accordance with the well-known P-FIT, according to which fronto-parietal regions represent the main neural substrates of intelligence (Jung and Haier, 2007), and with previous where fluid intelligence was associated to FC of fronto-parietal as well as temporal and occipital regions encompassing different resting-state networks such as the DAN and the ECN (Santarnecchi et al., 2017a, 2017b). In Rest2, however, the association between nv-F and p-GBC was found only in the lateral PFC, IPL, right temporal cortex, and cerebellum. While the peaks of this association were largely consistent with those found in Rest1 (i.e., lateral PFC and IPL), it is worth noting that the Penn Matrix Test was performed on the same day of the Rest1, but not Rest2, rs-fMRI session (Van Essen et al., 2013). It is possible that the reduced association between nv-F and p-GBC in Rest2 might be related to the different acquisition time of the Penn Matrix test as well as to fluctuations of the resting-state BOLD signal. Overall, the group of regions herein associated with nv-F is consistent with the pattern reported in previous works (Santarnecchi et al., 2017b). The high resemblance with resting-state networks such as DAN and ECN indicates that higher fluid intelligence might be related to the ability to focus on the task at hand while suppressing distractors on one hand, and to cognitive flexibility and online monitoring of cognitive resources on the other, as suggested previously (Santarnecchi et al., 2017b). Overall, our results suggest that these abilities might be reflected by increased GBC and aid to the validity of GBC as a marker of fluid intelligence.

4.6. Limitations and future perspectives

The sample adopted in this study included only healthy young adult participants, therefore these results are unable to provide a comprehensive picture of GBC over the entire lifespan. In addition, future works will benefit from a longitudinal approach, which will enable to explore the evolution and stability of GBC over time. Finally, given the role of the DMN observed in this study, GBC might be adopted in future studies to explore potential alterations in disorders where the disruption of this network is prominent, such as the Alzheimer’s disease (Dennis and Thompson, 2014).

5. Conclusions

In this study, we used rs-fMRI data retrieved from HCP to calculate GBC and analyze positive and negative correlations separately. First, we investigated the stability of GBC using ICC; then, we explored which brain areas show higher or lower GBC; finally, we examined the association between GBC and age, sex, and nv-F. N-GBC was less stable than p-GBC, possibly reflecting the more state-dependent nature of the negative functional couplings between brain regions. Consistently with prior findings, we found that DMN regions, insula, and visual areas were the brain regions with the highest GBC, while weaker FC strength was mainly localized in subcortical regions, temporal cortex, and cerebellum. Age was related to a general decrease in p-GBC, while males showed higher p-GBC than females. This latter result has never been reported before, and future studies are needed to replicate it. Finally, nv-F was related to increased p-GBC in fronto-parietal, occipital, and temporal regions, confirming the key role of these regions as neural substrates of intelligence.

Fig. 6.

Fig. 6

Open in a new tab

Effect of nonverbal fluid intelligence (nv-F), measured as the number of correct answers at the Penn Matrix Test, on the positive global brain connectivity (GBC) computed from (a) Rest1 (averaged across the first LR and the RL scans) and (b) Rest2 (averaged across the second LR and the RL scans) sessions. Color bar is based on Z-values. Darker colors indicate higher Z-values. Significant results were identified by thresholding Z-values with a false-discovery rate (FDR) corrected p-value < 0.05 and using a cluster threshold > 200 voxels. Upper limit of the color bar is set as the highest Z-value across all analyses. Hot colors indicate a positive relationship between nv-F and GBC

Highlights.

  • Global brain connectivity (GBC) reflects brain’s most connected regions

  • Positive GBC reflects more stable couplings between brain regions

  • Higher age is related to lower positive GBC

  • Males show higher positive GBC

  • Fluid intelligence is related to higher positive GBC in fronto-parietal regions

Acknowledgments

Both imaging and behavior data were provided by the Human Connectome Project, WU-Minn Consortium (Principal Investigators: David Van Essen and Kamil Ugurbil; 1U54MH091657) funded by the 16 NIH Institutes and Centers that support the NIH Blueprint for Neuroscience Research; and by the McDonnell Center for Systems Neuroscience at Washington University in St. Louis. Research reported in this publication was supported by the National Institute On Aging of the National Institutes of Health under Award Number U01AG052564 and by funds provided by the McDonnell Center for Systems Neuroscience at Washington University in St. Louis. The HCP-Aging 2.0 Release data used in this report came from DOI: 10.15154/1520707.

Data and/or research tools used in the preparation of this manuscript were obtained from the National Institute of Mental Health (NIMH) Data Archive (NDA). NDA is a collaborative informatics system created by the National Institutes of Health to provide a national resource to support and accelerate research in mental health. Dataset identifier(s): [NIMH Data Archive Digital Object Identifier 10.15154/1526336]. This manuscript reflects the views of the authors and may not reflect the opinions or views of the NIH or of the Submitters submitting original data to NDA. The authors thank the Human Connectome Project for open access to its data.

Funding

This work was supported by the National Institute on Aging [R21AG082345, R01AG060054, R01AG070227]; the National Institute of Biomedical Imaging and Bioengineering [R01EB031080-01A1, P41EB029460-01A1]; and the University of Maryland Baltimore, Institute for Clinical & Translational Research (ICTR) [1UL1TR003098].

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declarations of interest

None.

Data availability

Data are available under the approval of the Human Connectome Project, WU-Minn Consortium.

References

  1. Abraham A, Pedregosa F, Eickenberg M, Gervais P, Mueller A, Kossaifi J, Gramfort A, Thirion B, Varoquaux G, 2014. Machine learning for neuroimaging with scikit-learn. Front Neuroinform 8. 10.3389/fninf.2014.00014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bagarinao E, Watanabe H, Maesawa S, Mori D, Hara K, Kawabata K, Ohdake R, Masuda M, Ogura A, Kato T, Koyama S, Katsuno M, Wakabayashi T, Kuzuya M, Hoshiyama M, Isoda H, Naganawa S, Ozaki N, Sobue G, 2020. Identifying the brain’s connector hubs at the voxel level using functional connectivity overlap ratio. Neuroimage 222, 117241. 10.1016/j.neuroimage.2020.117241 [DOI] [PubMed] [Google Scholar]
  3. Barch DM, Burgess GC, Harms MP, Petersen SE, Schlaggar BL, Corbetta M, Glasser MF, Curtiss S, Dixit S, Feldt C, Nolan D, Bryant E, Hartley T, Footer O, Bjork JM, Poldrack R, Smith S, Johansen-Berg H, Snyder AZ, Van Essen DC, 2013. Function in the human connectome: Task-fMRI and individual differences in behavior. Neuroimage 80, 169–189. 10.1016/j.neuroimage.2013.05.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bilker WB, Hansen JA, Brensinger CM, Richard J, Gur RE, Gur RC, 2012. Development of Abbreviated Nine-Item Forms of the Raven’s Standard Progressive Matrices Test. Assessment 19, 354–369. 10.1177/1073191112446655 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Biswal B, Zerrin Yetkin F, Haughton VM, Hyde JS, 1995. Functional connectivity in the motor cortex of resting human brain using echo-planar mri. Magn Reson Med 34, 537–541. 10.1002/mrm.1910340409 [DOI] [PubMed] [Google Scholar]
  6. Braun U, Plichta MM, Esslinger C, Sauer C, Haddad L, Grimm O, Mier D, Mohnke S, Heinz A, Erk S, Walter H, Seiferth N, Kirsch P, Meyer-Lindenberg A, 2012. Test–retest reliability of resting-state connectivity network characteristics using fMRI and graph theoretical measures. Neuroimage 59, 1404–1412. 10.1016/j.neuroimage.2011.08.044 [DOI] [PubMed] [Google Scholar]
  7. Bullmore E, Sporns O, 2009. Complex brain networks: graph theoretical analysis of structural and functional systems. Nat Rev Neurosci 10, 186–198. 10.1038/nrn2575 [DOI] [PubMed] [Google Scholar]
  8. Camargo A, Del Mauro G, Wang Z, 2024. Task-induced changes in brain entropy. J Neurosci Res 102. 10.1002/jnr.25310 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cattell RB, 1963. Theory of fluid and crystallized intelligence: A critical experiment. J Educ Psychol 54, 1–22. 10.1037/h0046743 [DOI] [PubMed] [Google Scholar]
  10. Chang C, Glover GH, 2010. Time–frequency dynamics of resting-state brain connectivity measured with fMRI. Neuroimage 50, 81–98. 10.1016/j.neuroimage.2009.12.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Chen B, Xu T, Zhou C, Wang L, Yang N, Wang Z, Dong H-M, Yang Z, Zang Y-F, Zuo X-N, Weng X-C, 2015. Individual Variability and Test-Retest Reliability Revealed by Ten Repeated Resting-State Brain Scans over One Month. PLoS One 10, e0144963. 10.1371/journal.pone.0144963 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cole MW, Ito T, Braver TS, 2015. Lateral Prefrontal Cortex Contributes to Fluid Intelligence Through Multinetwork Connectivity. Brain Connect 5, 497–504. 10.1089/brain.2015.0357 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Cole MW, Pathak S, Schneider W, 2010. Identifying the brain’s most globally connected regions. Neuroimage 49, 3132–3148. 10.1016/j.neuroimage.2009.11.001 [DOI] [PubMed] [Google Scholar]
  14. Cole MW, Yarkoni T, Repovs G, Anticevic A, Braver TS, 2012. Global Connectivity of Prefrontal Cortex Predicts Cognitive Control and Intelligence. Journal of Neuroscience 32, 8988–8999. 10.1523/JNEUROSCI.0536-12.2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Del Mauro G, Wang Z, 2024. rsfMRI-based Brain Entropy is negatively correlated with Gray Matter Volume and Surface Area. bioRxiv 2024.04.28.591371. 10.1101/2024.04.28.591371 [DOI] [PubMed] [Google Scholar]
  16. Del Mauro G, Wang Z, 2023. Associations of Brain Entropy Estimated by Resting State fMRI With Physiological Indices, Body Mass Index, and Cognition. Journal of Magnetic Resonance Imaging. 10.1002/jmri.28948 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Del Mauro G, Sevel LS, Boissoneault J, Wang Z, 2024. Divergent association between pain intensity and resting-state fMRI-based brain entropy in different age groups. J Neurosci Res 102. 10.1002/jnr.25341 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Dennis EL, Thompson PM, 2014. Functional Brain Connectivity Using fMRI in Aging and Alzheimer’s Disease. Neuropsychol Rev 24, 49–62. 10.1007/s11065-014-9249-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Glasser MF, Sotiropoulos SN, Wilson JA, Coalson TS, Fischl B, Andersson JL, Xu J, Jbabdi S, Webster M, Polimeni JR, Van Essen DC, Jenkinson M, 2013. The minimal preprocessing pipelines for the Human Connectome Project. Neuroimage 80, 105–124. 10.1016/j.neuroimage.2013.04.127 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gopinath K, Krishnamurthy V, Cabanban R, Crosson BA, 2015. Hubs of Anticorrelation in High-Resolution Resting-State Functional Connectivity Network Architecture. Brain Connect 5, 267–275. 10.1089/brain.2014.0323 [DOI] [PubMed] [Google Scholar]
  21. Jung RE, Haier RJ, 2007. The Parieto-Frontal Integration Theory (P-FIT) of intelligence: Converging neuroimaging evidence. Behavioral and Brain Sciences 30, 135–154. 10.1017/S0140525X07001185 [DOI] [PubMed] [Google Scholar]
  22. Li Z, Kadivar A, Pluta J, Dunlop J, Wang Z, 2012. Test–retest stability analysis of resting brain activity revealed by blood oxygen level-dependent functional MRI. Journal of Magnetic Resonance Imaging 36, 344–354. 10.1002/jmri.23670 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Liang X, Zou Q, He Y, Yang Y, 2013. Coupling of functional connectivity and regional cerebral blood flow reveals a physiological basis for network hubs of the human brain. Proceedings of the National Academy of Sciences 110, 1929–1934. 10.1073/pnas.1214900110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lin L, Chang D, Song D, Li Y, Wang Z, 2022. Lower resting brain entropy is associated with stronger task activation and deactivation. Neuroimage 249, 118875. 10.1016/j.neuroimage.2022.118875 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lv H, Wang Z, Tong E, Williams LM, Zaharchuk G, Zeineh M, Goldstein-Piekarski AN, Ball TM, Liao C, Wintermark M, 2018. Resting-State Functional MRI: Everything That Nonexperts Have Always Wanted to Know. American Journal of Neuroradiology. 10.3174/ajnr.A5527 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Marek S, Hwang K, Foran W, Hallquist MN, Luna B, 2015. The Contribution of Network Organization and Integration to the Development of Cognitive Control. PLoS Biol 13, e1002328. 10.1371/journal.pbio.1002328 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Moeller S, Yacoub E, Olman CA, Auerbach E, Strupp J, Harel N, Uğurbil K, 2010. Multiband multislice GE-EPI at 7 tesla, with 16-fold acceleration using partial parallel imaging with application to high spatial and temporal whole-brain fMRI. Magn Reson Med 63, 1144–1153. 10.1002/mrm.22361 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Moore TM, Reise SP, Gur RE, Hakonarson H, Gur RC, 2015. Psychometric properties of the Penn Computerized Neurocognitive Battery. Neuropsychology 29, 235–246. 10.1037/neu0000093 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Noble S, Scheinost D, Constable RT, 2019. A decade of test-retest reliability of functional connectivity: A systematic review and meta-analysis. Neuroimage 203, 116157. 10.1016/j.neuroimage.2019.116157 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Noble S, Spann MN, Tokoglu F, Shen X, Constable RT, Scheinost D, 2017. Influences on the Test–Retest Reliability of Functional Connectivity MRI and its Relationship with Behavioral Utility. Cerebral Cortex 27, 5415–5429. 10.1093/cercor/bhx230 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Parente F, Colosimo A, 2020. Functional connections between and within brain subnetworks under resting-state. Sci Rep 10, 3438. 10.1038/s41598-020-60406-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Raichle ME, 2015. The Brain’s Default Mode Network. Annu Rev Neurosci 38, 433–447. 10.1146/annurev-neuro-071013-014030 [DOI] [PubMed] [Google Scholar]
  33. Raven J, 2000. The Raven’s Progressive Matrices: Change and Stability over Culture and Time. Cogn Psychol 41, 1–48. 10.1006/cogp.1999.0735 [DOI] [PubMed] [Google Scholar]
  34. Reineberg AE, Gustavson DE, Benca C, Banich MT, Friedman NP, 2018. The Relationship Between Resting State Network Connectivity and Individual Differences in Executive Functions. Front Psychol 9. 10.3389/fpsyg.2018.01600 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Salvador R, Verdolini N, Garcia-Ruiz B, Jiménez E, Sarró S, Vilella E, Vieta E, Canales-Rodríguez EJ, Pomarol-Clotet E, Voineskos AN, 2020. Multivariate Brain Functional Connectivity Through Regularized Estimators. Front Neurosci 14. 10.3389/fnins.2020.569540 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Santarnecchi E, Emmendorfer A, Pascual-Leone A, 2017a. Dissecting the parieto-frontal correlates of fluid intelligence: A comprehensive ALE meta-analysis study. Intelligence 63, 9–28. 10.1016/j.intell.2017.04.008 [DOI] [Google Scholar]
  37. Santarnecchi E, Emmendorfer A, Tadayon S, Rossi S, Rossi A, Pascual-Leone A, 2017b. Network connectivity correlates of variability in fluid intelligence performance. Intelligence 65, 35–47. 10.1016/j.intell.2017.10.002 [DOI] [Google Scholar]
  38. Satterthwaite TD, Wolf DH, Roalf DR, Ruparel K, Erus G, Vandekar S, Gennatas ED, Elliott MA, Smith A, Hakonarson H, Verma R, Davatzikos C, Gur RE, Gur RC, 2015. Linked Sex Differences in Cognition and Functional Connectivity in Youth. Cerebral Cortex 25, 2383–2394. 10.1093/cercor/bhu036 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Shah LM, Cramer JA, Ferguson MA, Birn RM, Anderson JS, 2016. Reliability and reproducibility of individual differences in functional connectivity acquired during task and resting state. Brain Behav 6. 10.1002/brb3.456 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Shehzad Z, Kelly AMC, Reiss PT, Gee DG, Gotimer K, Uddin LQ, Lee SH, Margulies DS, Roy AK, Biswal BB, Petkova E, Castellanos FX, Milham MP, 2009. The Resting Brain: Unconstrained yet Reliable. Cerebral Cortex 19, 2209–2229. 10.1093/cercor/bhn256 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Siman-Tov T, Bosak N, Sprecher E, Paz R, Eran A, Aharon-Peretz J, Kahn I, 2017. Early Age-Related Functional Connectivity Decline in High-Order Cognitive Networks. Front Aging Neurosci 8. 10.3389/fnagi.2016.00330 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Smith SM, Beckmann CF, Andersson J, Auerbach EJ, Bijsterbosch J, Douaud G, Duff E, Feinberg DA, Griffanti L, Harms MP, Kelly M, Laumann T, Miller KL, Moeller S, Petersen S, Power J, Salimi-Khorshidi G, Snyder AZ, Vu AT, Woolrich MW, Xu J, Yacoub E, Uğurbil K, Van Essen DC, Glasser MF, 2013. Resting-state fMRI in the Human Connectome Project. Neuroimage 80, 144–168. 10.1016/j.neuroimage.2013.05.039 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sporns O, Honey CJ, Kötter R, 2007. Identification and Classification of Hubs in Brain Networks. PLoS One 2, e1049. 10.1371/journal.pone.0001049 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Tadayon E, Pascual-Leone A, Santarnecchi E, 2020. Differential Contribution of Cortical Thickness, Surface Area, and Gyrification to Fluid and Crystallized Intelligence. Cerebral Cortex 30, 215–225. 10.1093/cercor/bhz082 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tomasi D, Shokri-Kojori E, Volkow ND, 2016. High-Resolution Functional Connectivity Density: Hub Locations, Sensitivity, Specificity, Reproducibility, and Reliability. Cerebral Cortex 26, 3249–3259. 10.1093/cercor/bhv171 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Tomasi D, Volkow ND, 2012. Gender differences in brain functional connectivity density. Hum Brain Mapp 33, 849–860. 10.1002/hbm.21252 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Tomasi D, Volkow ND, 2010. Functional connectivity density mapping. Proceedings of the National Academy of Sciences 107, 9885–9890. 10.1073/pnas.1001414107 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Tomasi D, Wang G-J, Volkow ND, 2013. Energetic cost of brain functional connectivity. Proceedings of the National Academy of Sciences 110, 13642–13647. 10.1073/pnas.1303346110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. van den Heuvel MP, Sporns O, 2013. Network hubs in the human brain. Trends Cogn Sci 17, 683–696. 10.1016/j.tics.2013.09.012 [DOI] [PubMed] [Google Scholar]
  50. Van Essen DC, Smith SM, Barch DM, Behrens TEJ, Yacoub E, Ugurbil K, 2013. The WU-Minn Human Connectome Project: An overview. Neuroimage 80, 62–79. 10.1016/j.neuroimage.2013.05.041 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Varangis E, Habeck CG, Razlighi QR, Stern Y, 2019. The Effect of Aging on Resting State Connectivity of Predefined Networks in the Brain. Front Aging Neurosci 11. 10.3389/fnagi.2019.00234 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Wang Z, 2021. The neurocognitive correlates of brain entropy estimated by resting state fMRI. Neuroimage 232, 117893. 10.1016/j.neuroimage.2021.117893 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wang Z, Li Z, Pluta J, Detre JA, 2012. Improving fMRI activation detection sensitivity using intervoxel coherence mapping. Int J Imaging Syst Technol 22, 33–36. 10.1002/ima.21297 [DOI] [Google Scholar]
  54. Zuo X-N, Ehmke R, Mennes M, Imperati D, Castellanos FX, Sporns O, Milham MP, 2012. Network Centrality in the Human Functional Connectome. Cerebral Cortex 22, 1862–1875. 10.1093/cercor/bhr269 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Data are available under the approval of the Human Connectome Project, WU-Minn Consortium.

RESOURCES