Next Article in Journal
Thirty Risk Factors for Alzheimer’s Disease Unified by a Common Neuroimmune–Neuroinflammation Mechanism
Previous Article in Journal
Familiarity Processing through Faces and Names: Insights from Multivoxel Pattern Analysis
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Brain Sciences
Volume 14
Issue 1
10.3390/brainsci14010040
brainsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Abnormal Spatial and Temporal Overlap of Time-Varying Brain Functional Networks in Patients with Schizophrenia

by
Jie Xiang
1,
Yumeng Sun
1,
Xubin Wu
1,
Yuxiang Guo
2,
Jiayue Xue
1,
Yan Niu
1 and
Xiaohong Cui
1,*
1
College of Computer Science and Technology, Taiyuan University of Technology, Taiyuan 030024, China
2
School of Software, Taiyuan University of Technology, Taiyuan 030024, China
*
Author to whom correspondence should be addressed.
Brain Sci. 2024, 14(1), 40; https://doi.org/10.3390/brainsci14010040
Submission received: 11 December 2023 / Revised: 25 December 2023 / Accepted: 27 December 2023 / Published: 31 December 2023
(This article belongs to the Section Psychiatric Diseases)
Browse Figures
Versions Notes

Abstract

:
Schizophrenia (SZ) is a complex psychiatric disorder with unclear etiology and pathological features. Neuroscientists are increasingly proposing that schizophrenia is an abnormality in the dynamic organization of brain networks. Previous studies have found that the dynamic brain networks of people with SZ are abnormal in both space and time. However, little is known about the interactions and overlaps between hubs of the brain underlying spatiotemporal dynamics. In this study, we aimed to investigate different patterns of spatial and temporal overlap of hubs between SZ patients and healthy individuals. Specifically, we obtained resting-state functional magnetic resonance imaging data from the public dataset for 43 SZ patients and 49 healthy individuals. We derived a representation of time-varying functional connectivity using the Jackknife Correlation (JC) method. We employed the Betweenness Centrality (BC) method to identify the hubs of the brain’s functional connectivity network. We then applied measures of temporal overlap, spatial overlap, and hierarchical clustering to investigate differences in the organization of brain hubs between SZ patients and healthy controls. Our findings suggest significant differences between SZ patients and healthy controls at the whole-brain and subnetwork levels. Furthermore, spatial overlap and hierarchical clustering analysis showed that quasi-periodic patterns were disrupted in SZ patients. Analyses of temporal overlap revealed abnormal pairwise engagement preferences in the hubs of SZ patients. These results provide new insights into the dynamic characteristics of the network organization of the SZ brain.

1. Introduction

Schizophrenia is a serious, chronic mental illness that often develops in early adulthood [1]. Hallucinations, delusions, speech and behavioral disturbances, and mood disorders are the hallmark symptoms of this psychiatric disorder [2]. Neuroimaging studies have facilitated extensive research into schizophrenia, and complex network analysis methods are frequently utilized in this field. Research has shown that schizophrenia is a complex psychiatric disorder characterized by abnormalities in whole-brain connectivity and is associated with abnormal functional coordination between multiple brain regions.
With the advancement of neuroscience tools, various neuroimaging methods, such as electroencephalography (EEG)/intracranial electroencephalography (iEEG), functional magnetic resonance imaging (fMRI), magnetoencephalography (MEG), and event-related potentials (ERPs), have been employed to investigate brain function [3,4,5,6]. Notably, functional magnetic resonance imaging (fMRI) stands out due to its exceptional spatial resolution, enabling precise anatomical localization and visualization of deep brain structures [7]. This method offers comprehensive whole-brain coverage, facilitating the exploration of functional connectivity and network-level brain function, distinguishing it from other techniques that may have limitations in capturing activity across the entire brain simultaneously [8]. Furthermore, fMRI’s capacity to measure blood flow changes provides insights into hemodynamic responses associated with neural activity, aiding in the identification of anatomical correlates of cognitive function [9]. Resting-state functional magnetic resonance imaging (rs-fMRI) has become an essential tool for exploring the functional networks of the human brain. Traditional methods have assumed stable functional connectivity throughout scans when constructing brain networks. However, recent studies have shown that the human brain is dynamic, displaying scale-free fluctuations in space and time [10]. As such, dynamic network analysis methods have been increasingly adopted to better detect fluctuations in functional brain connectivity [11,12]. For instance, Cui et al. utilized dynamic functional connectivity to capture changes in network topology and cognitive behavior over time, providing a better understanding of the time-varying characteristics of brain networks [13]. Similarly, Braun et al. combined pharmacological interventions with novel techniques in dynamic network neuroscience to study alterations in the dynamic remodeling of brain networks associated with genetic risks for schizophrenia [14]. Meanwhile, Zhu et al. used sliding windows and Pearson correlation to construct dynamic brain networks to study functional connectivity abnormalities and network disruptions in the whole brain [15]. Additionally, Gifford et al. introduced a new method for measuring large-scale dynamic brain organization called dynamic modular organization. They used it to examine intergroup differences in dynamic community structure in SZ [16]. These studies collectively underscore the growing recognition that schizophrenia is a disorder caused by abnormalities in the dynamic organization of brain networks [17,18,19].
Previous studies have focused on changes in functional connectivity within brain networks and changes in specific brain communities. However, graph theory-based network analyses have shown that the presence of hubs represents a vital feature of the organization of large-scale brain networks. As highly connected regions, hubs integrate and distribute information throughout the brain and facilitate effective communication between different brain regions. Despite this, previous investigations have neglected to examine hubs’ dynamic reorganization and spatiotemporal patterns. Our study, therefore, aims to fill this gap by examining the dynamic spatiotemporal organization patterns of hubs in patients diagnosed with schizophrenia through temporal overlap, spatial overlap, and hierarchical clustering [20,21].
Here, we aimed to investigate abnormal changes in the dynamic spatiotemporal organization of brain hubs in SZ patients at rest, thus providing a new characterization of the dynamic network in SZ patients. Specifically, we investigate the spatiotemporal organization of hubs in dynamic brain networks based on a time-varying approach to functional connectivity. A time-varying network is established using the Jackknife Correlation (JC) method to obtain the dynamic functional connectivity of the network, and the Betweenness Centrality (BC) method is used to identify the hubs in the brain. Temporal overlap, spatial overlap, and hierarchical clustering are then used to quantify fluctuations in the brain hubs.

2. Materials and Methods

2.1. Participants

The data were selected from the UCLA Consortium for Neuropsychiatric Phenomics LA5c Study project, which was approved by a UCLA Institutional Review Board. These data were obtained through a public database, open fMRI (https://openfmri.org/dataset/ds000030/, accessed on 1 September 2020). Our study included 92 participants, including 43 people with schizophrenia (SZ) and 49 normal controls (NC). There were no significant differences in age and sex between NC and SZ (t-test). It is worth noting that 43 SZ patients out of 92 subjects participated in the SAPS assessment. The SAPS (Scale for the Assessment of Positive Symptoms) is a clinical rating scale used to assess positive symptoms of schizophrenia, including hallucinations, delusions, and disorganized thinking. Specific demographic characteristics are shown in Table 1.

2.2. Imaging Acquisition and Preprocessing

All subjects underwent MRI scans with 32-channel head coils on a 3-T scanner at the University of Electronic Science and Technology of China. Resting-state fMRI data were collected using a series of T2*-weighted echo plane imaging (EPI). The parameters are as follows: Repeat Time (TR) = 2 s, Echo Time (TE) = 30 ms, Slice Thickness = 4 mm, Slices = 34, Flip Angle = 90°, Field of View (FOV) = 192 mm, Matrix = 64 × 64. The resting-state fMRI scan lasted a total of 304 s. For the structure scan, a T1-weighted high-resolution anatomical scan was performed using the following parameters: slice thickness = 1 mm, slices = 176, repeat time (TR) = 1.9 s, echo time (TE) = 2.26 ms, matrix = 256 × 256, field of view (FOV) = 250 mm.
Data preprocessing was carried out using the DPABI toolbox (a toolbox for Data Processing & Analysis for Brain Imaging, V2.1, http://rfmri.org/dpabi, accessed on 30 July 2023), the first 10 volumes of signals were discarded, and data for the remaining 142 time points were preprocessed as follows: (1) Slice timing correction. (2) Realignment. (3) Normalization: Image space normalized to the Montreal Neurological Institute (MNI) head anatomy template and resampled with 3 × 3 × 3 mm3 voxels [22]. (4) Filtering: bandpass filtering (0.01 ≤ f ≤ 0.1 Hz) was performed on the image. (5) Smoothing: the images were spatially smoothed using a Gaussian filter with a full-width at half-maximum (FWHM) of 6 mm [23]. (6) Remove the covariates and divide the brain into 90 regions using the Automated Anatomical Marker Template (AAL) [24], with each voxel extracting the residual time series.
Based on the time series extracted from the AAL template (Detailed information is provided in Table A1), we divided brain regions into 5 functional networks [25]: somatosensory/motor and auditory network (SMN), visual network (VN), attention network (AN), default mode network (DMN), and limbic/paralimbic and subcortical network (LSN).

2.3. Calculation of Time-Varying Functional Connectivity (TVC)

We used the Jackknife Correlation (JC) method to evaluate functional connectivity over time [26]. The JC of two BOLD time series x and y at time point t can be expressed as
J C t = − Σ t i t t o t a l x t i − x t ¯ y t i − y t ¯ Σ t i t t o t a l x t i − x t ¯ 2 Σ t i t t o t a l y t i − y t ¯ 2 ,   t i ≠ t
where t t o t a l is the total number of time points. x t ¯ and y t ¯ are the expected values excluding the data at time point t :
x t ¯ = 1 t t o t a l − 1 ∑ t i t t o t a l x t i ,   t i ≠ t
y t ¯ = 1 t t o t a l − 1 ∑ t i t t o t a l y t i ,   t i ≠ t
Notably, previous studies have shown that the JC method performs better in tracking covariance time changes than other methods. Here, each JC value is normalized, and the standardized JC method is unaffected by the underlying static properties. The JC method estimates paired functional brain connections at each time point. This information can be used to investigate the dynamic characteristics of a given network and the interactions between different nodes within it.

2.4. Calculation of Betweenness Centrality at Multilevel

In order to identify candidate hubs and active hubs in the network, it is necessary to calculate the Betweenness Centrality (BC) [27,28] value for each node using a graph measure. This metric is a basis for determining whether a given node is a hub. BC is a type of network centrality metric that quantifies the influence of a particular node in connecting other nodes within the network. Nodes with high BC values are often called “hubs” or “high centers”. These nodes are critical in inter-node communication, facilitating information transfer between other nodes. The normalized BC of node n i is defined as
B C n i = 2 n t o t a l − 1 n t o t a l − 2 ∑ n h ,   n j ∈ N n h ≠ n j ,   n h ≠ n i ,   n j ≠ n i ρ n h n j n i ρ n h n j
where n t o t a l is the total number of all nodes, N is the set containing all nodes, ρ n h n j is the number of shortest paths between n h and n j , and ρ n h n j n i is the number of shortest paths between node n h and node n j that pass through node n i .
Meanwhile, we calculated the time average of BC for node n i as
B C _ a v g n i = 1 t t o t a l Σ t = 1 t t o t a l B C n i t
After calculating the BC values at the node level, to further analyze our results, we studied the BC values at the global level and the resting-state network (RSN) level. By examining the BC values at different levels, we can better understand the network structure. Precisely, the BC values at each level were calculated using the following formulas:
Global BC: The sum of the BC values of all nodes divided by the total number of nodes in the network. Global BC is defined as
B C _ a v g g l o b a l = 1 n t o t a l Σ n i = 1 n t o t a l B C _ a v g n i
RSN BC: The average BC value of nodes belonging to a given RSN in the network. BC for R S N j is defined as
B C _ a v g R S N j = 1 n R S N j ∑ n i ∈ N R S N j B C _ a v g n i
where n R S N j is the total number of nodes in R S N j , n i is the set of all nodes contained in R S N j , N R S N j is the set of nodes in R S N j .

2.5. Identification of Candidate Hubs and Active Hubs

A node is considered a candidate hub if it exhibits a higher time-averaged Betweenness Centrality (BC) value than other nodes. To identify candidate hubs, a threshold value of one standard deviation above the mean BC value is used, as represented by T H 1 [29]. Nodes with BC values above this threshold are selected as candidate hubs, allowing for targeted analysis of high-impact nodes within the network.
T H 1 = 1 n t o t a l Σ n i = 1 n t o t a l B C _ a v g n i + Σ n i = 1 n t o t a l ( B C _ a v g n i − 1 n t o t a l Σ n i = 1 n t o t a l B C _ a v g n i ) 2 n t o t a l − 1
Our study recognizes that hubs are not static and can change over time. Thus, we examined candidate hubs from a time-varying perspective. Among the candidate hubs, we define those with larger time-varying BC values as active hubs. In other words, we focused on active hubs with time-varying properties. Active hubs are a subset of candidate hubs with larger BC values at specific time points. A separate threshold is applied to the time-varying BC values of candidate hubs to identify active hubs, as represented by T H 2 [29]. Specifically, an active hub is defined as a candidate hub whose BC value exceeds one standard deviation of its average BC value. By identifying and analyzing active hubs, researchers can gain insight into each node’s dynamic role in the network’s overall connectivity structure over time.
T H 2 = B C _ a v g n i + Σ t = 1 t t o t a l ( B C n i t − B C _ a v g n i ) 2 t t o t a l − 1

2.6. Subnetwork Distribution of Active Hubs

In this section, our study aimed to investigate the properties of active hubs in different brain subnetworks. In other words, the distribution of active hubs in different subnetworks was studied. The evidence suggests that the number of active hubs fluctuates over time and exhibits different subnetwork distributions for different subnetworks [30]. To investigate whether the spatial configuration of active hubs differed within any particular subnetwork, we calculated the relative number of active hubs in each subnetwork for all subjects at all points in time. This approach allowed us to determine whether any particular subnetwork had more active hubs than others. The subnetwork distribution of active hubs (SND) belonging to a particular R S N j in the network is defined as:
S N D R S N j = ∑ n j ∈ C ∩ n j ∈ N R S N j ,   t ∈ T A n j ( t ) ∑ n i ∈ C ,   t ∈ T A n i ( t )
where A n i ( t ) = 1 if node n i is an active hub at time t . C is the set of candidate hubs. N R S N j is the set of nodes in R S N j . T is the set of all time points.

2.7. Spatial Overlap and Temporal Overlap of Active Hubs

2.7.1. Spatial Overlap of Active Hubs

Spatial overlap (SO) measures the spatial patterns of brain activity and provides valuable insight into the quasi-periodic nature of brain function. Specifically, SO involves the identification of reliable patterns of brain activity that are repeated over time. Mathematically, the degree of SO at time points t i and t j is calculated as the ratio of the intersection of active hubs at the two time points divided by the union of active hubs at those two time points. In this way, the spatial overlap between active hubs at different time points can be precisely quantified, providing insight into brain connectivity dynamics. The SO at time points t i and t j is expressed as:
S O ( t i , t j ) = S t i ∩ S t j S t i ∪ S t j
where S t i is the set of active hubs at time point t i , and S t j is the set of active hubs at time point t j .

2.7.2. Hierarchical Clustering of Active Hubs

To explore the quasi-periodic patterns of brain space configurations, we utilized a top-down hierarchical clustering algorithm [31] to cluster these patterns in a rational way. In this approach, each pattern represents the set of active hubs at each time point, and we used spatial overlap between patterns as a similarity metric. We used the average linkage method to calculate the cluster-to-cluster distance. The results following hierarchical clustering provide insight into the quasi-periodic patterns of the spatial configuration of the brain in SZ and NC.

2.7.3. Temporal Overlap of Active Hubs

Temporal overlap (TO) measures the interaction of the brain hubs’ activity, providing insight into the frequency patterns at which pairs of active hubs co-occur. In other words, TO represents the proportion of time during which paired active hubs are both active. By examining the degree of TO between hubs, we can gain essential insights into the pairwise participation of these hubs. Mathematically, the TO between compute nodes n i and n j can be expressed as the proportion of time during which both nodes are active hubs. This allows for the precise quantification of the degree of temporal overlap between candidate hubs, which can be used to identify hubs that exhibit particularly strong or consistent patterns of pairwise activity. By analyzing the temporal overlap of active hubs across different time points, researchers can gain insight into the temporal dynamics of activity within the brain network under study. The time overlap of compute nodes n i and n j is expressed as:
T O ( n i , n j ) = S n i ∩ S n j t t o t a l
where S n i is the set of time points where node n i is an active hub, and S n j is the set of time points where node n j is an active hub and t t o t a l is the number of total time points.

2.8. Statistical Analysis

An independent sample t-test was used for statistical tests to quantify the differences between NC and SZ. For each metric, we averaged the metrics obtained from the network. The Benjamini and Hochberg error discovery rate (BH_FDR) method was used to calibrate all the results, and the threshold value for the significant difference was set to <0.05. The pipeline of the analysis strategy for this study is displayed in Figure 1.

3. Results

3.1. Group Comparisons on BC Values

This study aimed to identify hubs by using BC values and assess the differences in BC between the two groups of subjects. We analyzed these differences at different levels, which included the whole brain and the resting state network (RSN). At the whole-brain level, our results showed that the BC values of SZ patients were significantly lower than those of NC subjects (p < 0.01), as shown in Figure 2A. At the RSN level, when using FDR correction to control for multiple comparisons, we found significantly higher BC in the default mode network (DMN) (p (FDR) < 0.05) and significantly lower BC in the attentional network (AN) (p (FDR) < 0.05) and visual network (VN) (p (FDR) < 0.05) in SZ patients compared to the NC group (Figure 2B).

3.2. Candidate Hubs

Our study also identified candidate hubs associated with larger BC values, which are shown in Figure 3. We found significant differences in the anatomical positioning of candidate hubs between SZ patients and NC subjects. In the NC group, we identified a total of 14 candidate hubs, which were distributed among DMN, AN, SMN, LSN, and VN. Specifically, there were three candidate hubs in DMN, four in AN, three in SMN, two in LSN, and two in VN. In contrast, in the SZ group, we identified 14 candidate hubs. These candidate hubs consisted of five in DMN, two in AN, two in SMN, and five in LSN. Of note, seven candidate nodes were consistent between the two groups.

3.3. Distribution of Active Hubs

We investigated the active hubs from the perspective of subnetwork distribution. It has been shown that the number of hubs fluctuates over time and varies for the distribution of different subnetworks in the brain. To explore whether the spatial configuration of active hubs differed in any system, we calculated the relative number of active hubs in each subnetwork at all time points for all subjects. The results are displayed in Figure 4. We can see that the relative proportion of active hubs in each subnetwork differs between NC and SZ groups. The most significant proportion of active hubs for the NC group is found in the AN, followed by DMN and SMN. The LSN and VN have the lowest proportion of active hubs. For the SZ group, the most significant proportion of active hubs is found in the DMN and LSN, followed by AN and SMN. The VN has the lowest proportion of active hubs in this group.

3.4. Spatial Overlap of Active Hubs

In this part of the study, we averaged the spatial overlap for each subject at all time points. As shown in Figure 5, after statistical analysis, the spatial overlap was lower in SZ patients than in NC ( p > 0.05).
Furthermore, seven clusters emerged in SZ patients after choosing the same threshold of 0.90 for the hierarchical clustering of brain space patterns. That is, the spatial patterns of the brain in SZ patients varied periodically between these seven clusters. However, NC was grouped into only three clusters and varied periodically between the three patterns. We found that the hierarchical clustering patterns of SZ patients were more dispersed (Figure 6C). In contrast, the quasi-cyclic patterns of NC were more stable (Figure 6A). The spatial patterns following the hierarchical clustering of NC are shown in Figure 6B. The spatial patterns following the hierarchical clustering of SZ patients are shown in Figure 6D. Notably, the SZ clustering contained seven spatial patterns, four of which were isolated.

3.5. Temporal Overlap of Active Hubs

To measure the temporal overlap of active hubs, we calculated the number of points in time that were jointly active on each pair of hubs as a proportion of the total points in time. The results of the temporal overlap are shown in Figure 7. Figure 7A shows the results of the temporal overlap of all hubs in NC. Figure 7B shows the results for SZ patients. Notably, SZ patients had the highest degree of co-occurrence of the REC.L and REC.R located in the DMN, and were also relatively high for the TPOmid.L and TPOmid.R located in LSN, with the remaining significant node pairs being between DMN and LSN. The highest degree of temporal overlap in NC occurred in ANG.L and ANG.R in AN, followed by IPL.R in AN with SMG.R in SMN, with other sizeable temporal overlap pairs in SMN or between SMN and AN. Figure 7C,D shows the distribution of the higher temporal overlap in nodal pairs for NC versus SZ patients. The thickness of the connecting lines in the graph represents the degree of time overlap. The top five most significant temporal overlap hub pairs for NC versus SZ patients are shown in Figure 7E,F.

4. Discussion

We used Jackknife Correlation to construct time-varying dynamic brain networks to investigate the spatial and temporal organization patterns of SZ brain hubs. This paper utilizes the BC method in graph theory to identify brain hubs. Meanwhile, the subnetwork distribution is presented, and the hubs are studied from the perspective of the subnetwork distribution. Further, the difference between SZ and NC is analyzed using temporal overlap, spatial overlap, and hierarchical clustering of hubs. The results showed a trend towards decreased repetition and periodicity in the spatial configuration of brain hub activity and abnormalities in hub pair participation preferences in SZ patients. Research on brain hubs based on fMRI studies can help identify abnormal brain activity or connectivity patterns associated with neurological and psychiatric disorders. Comparing brain activity in healthy individuals and those with disorders such as schizophrenia can help elucidate the neural basis of these disorders, potentially aiding early diagnosis, treatment and therapeutic intervention.

4.1. Decreased Ability of SZ Patients to Integrate and Process Information

BC represents the ability of a region to integrate information from other brain networks. Our study showed that BC was significantly reduced at the whole-brain level in patients with SZ (Figure 2A). This result suggests that the brain of SZ patients has a reduced ability to integrate and process information. A possible explanation for our findings could be a reduction in the patient’s brain’s global communication capacity. A study exploring the role in the global functioning of brain function in schizophrenia showed that the rich club organization between the hubs was significantly affected in SZ patients. Selective disruption of brain connectivity between the hubs in the patient’s brain may lead to reduced communication capacity and altered functional brain dynamics [32]. Furthermore, a study based on the dynamic reconfiguration of resting-state functional connectivity suggests a fundamental role for elevated informational communication instability and diminished whole-brain integration of higher-order networks in schizophrenia [33]. These correlational studies with our results provide neurological evidence to support the hypothesis of reduced global communication capacity in schizophrenia.
We have also made a set of comparisons of BC from the RSN level. The results showed significant differences in DMN, VN, and AN between SZ patients and NC (Figure 2B). Firstly, for DMN, our results found significantly higher BC in SZ patients than in NC, suggesting the most significant contribution of this network to the compensatory mechanisms caused by the patients’ disrupted information integration [34]. Studies have found that the DMN in SZ patients is frequently over-activated and over-connected [35], which may be associated with excessive self-reference and impairment in attention and working memory [36]. Based on these findings, our results suggest that abnormal DMN states affect the SZ patient’s cognitive function, leading to mental deficits. Secondly, for AN, our results found that SZ patients had significantly lower BC than NC, suggesting a prominent contribution of abnormalities in this network to reducing patients’ global information processing capacity. Abnormalities in attentional networks are closely associated with attentional deficits in schizophrenia [37,38]. Brain dysfunction resulting from abnormalities in the attentional network may lead to abnormal clinical and cognitive measures [39]. Finally, our results found significantly lower BC for VN in SZ patients. This finding further emphasizes the role of impaired visual processing in SZ. Alterations in the VN may contribute prominently to the observed deficits in information processing capacity in this patient population. Evidence suggests that schizophrenia is associated with deficits in higher-order visual information processing [40,41,42]. These findings suggest that abnormal visual network states are associated with SZ dysfunction. Furthermore, our conjecture is confirmed based on the results of the subnetwork distribution (Figure 4). Specifically, we observed a significantly higher proportion of subnetwork distribution in the DMN and a significantly lower proportion of active hubs in the AN and VN in SZ patients compared to NC, confirming the impairment of the AN and VN and the overactivation of the DMN in SZ.
In addition to the between-group differences in BC values that we observed at the whole-brain and subnetwork levels, there were also significant differences between the two groups in terms of candidate hubs (Figure 3). Specifically, when examining the inconsistent structure of candidate hubs in the SZ versus NC, we found that inconsistent candidate hubs in SZ patients were located in the DMN and LSN, potentially reflecting altered functional connectivity in these regions associated with SZ. These findings are consistent with previous studies that SZ is associated with regional differences in individual nodes of the DMN [43] and that hubs of SZ are more common in the LSN [44]. Our findings highlight differences in hub patterns between the SZ and NC and suggest that these differences may underlie observed differences in information processing capacity and cognitive function between the two groups.

4.2. Reduced Stability of Spatial Configuration of Brain Activity in SZ Patients

Recent research has found that the activities of the cerebral cortex are not separate. Instead, they overlap and interact at different scales, which may be attributable to dynamic fluctuations in human brain activity in space and time [45], and is associated with heterogeneous activity of mixed neurons within brain regions. Evidence from the study suggests that network overlap in the brain may provide clues to help understand network interactions in human cognition [46]. Describing the spatiotemporal organization of the brain and the differences between SZ patients and NC is also one of the central questions in the study of SZ as a psychiatric disorder [47]. We therefore combine this concept of overlap with spatiotemporal dynamics to investigate the dynamic interaction of abnormal spatiotemporal patterns in the brain hubs of SZ patients. Research reveals abnormal spatiotemporal patterns that indicate disrupted communication and coordination between brain regions in schizophrenia patients [48,49]. Understanding these spatiotemporal changes is critical to uncovering the neural mechanisms of schizophrenia and developing more targeted treatments or interventions [50,51].
Our study found that, on average, patients with SZ had lower spatial overlap than NC. Further, after hierarchical clustering based on the spatial overlap, we obtained the clustering patterns of quasi-periodic changes in the two groups of subjects. A quasi-periodic pattern is a recurring pattern of brain activity [52]. The clustering patterns were found to be more dispersed in the SZ patients, indicating an abnormal quasi-periodic pattern in the SZ patients [53]. These disordered quasi-periodic patterns may lead to abnormal functional connectivity in patients with SZ. This conjecture is supported by other psychiatric disorders, where disruption of the patient’s quasi-periodic patterns may result in a failure of the quasi-periodic patterns to maintain normal functional connectivity [54]. In contrast to NC’s spatial patterns after clustering, the altered hubs in SZ’s clustering patterns contain hubs such as TPOmid.L and TPOmid.R nodes. The bilateral TPOmid nodes are located in the temporal poles, the part of the temporal lobe that forms the most rostral part of the temporal lobe. The temporal pole is associated with higher cognitive processes and symptoms associated with temporal pole lesions or dysfunction can help identify schizophrenia [55]. In examining dynamic functional connectivity disorders associated with schizophrenia, Sun and colleagues found that patients with SZ had abnormal temporal area properties of the TPOmid bilaterally [56]. However, in contrast to NC, no ANG.R and ANG.L nodes located in AN were found in the post-clustering spatial patterns of the SZ patients. The bilateral ANG nodes are located in the posterior part of the lower parietal lobe and play a role in semantic processing, reading, and word comprehension [57]. The memory loss and slower cognitive speed exhibited by SZ patients may be related to the loss of ANG in their patterns. A study by Gao and colleagues also found that people with SZ showed reduced low-frequency fluctuations in the ANG [58]. Thus, there were abnormalities in the spatial patterns of SZ patients, and the changes in spatial patterns were more variable in SZ patients. These findings support the theory of spatial dynamics, whereby the spatial patterns of the brain change continuously over time. Furthermore, the spatial patterns of clustering reflect periodic changes in the spatial patterns of the brain. Research has shown a correlation between quasi-periodic patterns in the brain activity of individuals and global signals [52,59]. Global signals are common patterns of neural activity observed throughout the brain. The widespread abnormal global signals found in the SZ may lead to alterations in the SZ quasi-periodic pattern [60]. Here, we investigate the dynamics of brain abnormalities in SZ patients by examining quasi-periodic patterns in the brain. The quasi-periodic patterns provide insight into the mechanisms behind the differences in spatial patterns in SZ patients.

4.3. Abnormal Pair Participation Preference for Hubs in SZ Patients

Hub nodes play a key role in the brain’s functional brain connectivity network. They are important nodes that connect different brain regions and facilitate the dissemination and integration of information in the brain network. The activity of these hub nodes may increase when the brain has to process a large flow of information, creating a pattern of temporary, paired occurrences. These temporary hub connections provide a faster, more direct pathway to meet the demands of neuronal traffic. These temporary fast connections may facilitate the rapid transfer and processing of information. We use temporal overlap as a measure of the strength of such fast connections. Time overlap refers to the percentage of time that two hubs are active simultaneously. The temporal overlap reflects the temporal interaction pattern between hubs and indicates pair participation preference for hubs [61]. Overall, the temporal overlap was lower in SZ patients than in NC, suggesting that SZ patients have a reduced ability to interact with information. Our study shows that the hub pairs with the strongest participation preference in SZ patients are distributed between REC.L and REC.R. REC is the anterior frontal cortex of the anterior cingulate, which is implicated in the psychopathology of schizophrenia [62]. These regions are associated with olfactory function. A large proportion of SZ patients have olfactory hallucinations [63,64] and therefore prolonged abnormal activation of these regions may be associated with these olfactory abnormalities. Unusually, hubs’ highest pair participation preference was found between ANG.L and ANG.R in NC. Located in the posterior part of the brain [65], the ANG is part of the DMN and is involved in self-referential processing, imagery, and memory, playing an important role in language processes, spatial cognition, and memory retrieval [66]. The findings suggest that a control mechanism involving the ANG contributes to perceptual and cognitive problems in patients with SZ, which is essentially a neural communication problem, and that this control mechanism may be a worthy target for intervention [67]. Connectivity analysis has shown that abnormal information transfer between ANG areas in patients with SZ may be due to reduced connectivity between the ANG and frontal action-generating areas [68]. More specifically, ANG has become a cross-modal hub in which fused multisensory information is combined and integrated to allow it to make sense of events in a comprehensive and meaningful way and to redirect attention to relevant information. Furthermore, the dysfunction of ANG has been proposed in neurobiological models of social deficits and ego disorders observed in SZ patients [69]. In NC, brain nodes located in the ANG were active and were consistently active. Therefore, patients with SZ may have cognitive dysfunction due to abnormal pair participation preference for hubs in the ANG nodes. Our analysis also revealed that SZ patients showed an abnormal reduction in pair participation preference for SMG.L and SMG.R compared to NC. These SMG nodes are located in the parietal lobe, above the temporal lobe, in the vicinity of brain regions involved in language and hearing [70]. As a branch of the inferior parietal lobule, the SMG plays an important role in orientation regulation, proprioception, and emotion regulation [71], and this region has been repeatedly identified as abnormal in structural imaging studies of SZ patients [72]. SMG nodes play a crucial role in the processing of heard speech, and their abnormally reduced pattern of paired participation may contribute to the language-related deficits observed in SZ patients. These changes in pair participation preferences between hubs in the SZ represent a significant change in the rapid pathway of information transfer and processing of hubs.

5. Limitations and Directions for Future Research

This study had some limitations, mainly due to the small sample size. Although we obtained relatively stable and reliable results, the universality of the sample still needs to be explored and improved by recruiting more patients in future surveys. The study adopted a cross-sectional design, offering a snapshot of brain dynamics at a specific moment in time in individuals diagnosed with schizophrenia. However, to comprehensively grasp the dynamic nature of schizophrenia, future investigations employing longitudinal studies are imperative. Furthermore, in order to clarify the correlation and causality in the differences in brain network organization between patients with schizophrenia and healthy controls, future research directions should use longitudinal studies to track brain changes before and after the onset of illness.

6. Conclusions

This study employed a time-varying network model to investigate the spatial and temporal overlap of brain hubs in SZ patients. The findings revealed a notable decrease in Betweenness Centrality observed across multiple levels in SZ patients. Our results contribute substantiating evidence indicating that hubs in both healthy individuals (NC) and SZ patients undergo dynamic fluctuations even during resting states. However, hubs in SZ individuals demonstrate heightened spatial and temporal instability in their organization, manifesting frequent spatial reconfigurations and deviant pairwise engagement preferences. The identification of aberrant brain regions in this study offers promising avenues for further in-depth pathological investigations into SZ.

Author Contributions

Conceptualization, J.X. (Jie Xiang) and X.C.; methodology, X.C.; software, J.X. (Jiayue Xue) and Y.N.; validation, Y.S., X.W. and Y.G.; formal analysis, X.C. and Y.S.; investigation, X.C. and Y.S.; resources, X.C. and Y.S.; data curation, X.C. and Y.S.; writing—original draft preparation, X.C. and Y.S.; writing—review and editing, Y.S., X.W., Y.G., J.X. (Jiayue Xue), and Y.N.; visualization, Y.S. and X.W.; supervision, J.X. (Jie Xiang); project administration, X.C.; funding acquisition, J.X. (Jie Xiang). All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Functional of China (62376184, 62303445), China Postdoctoral Science Foundation (2023M733669), Shenzhen Basic Research Project (Natural Science Foundation) (JCYJ20230807140719040), Shanxi Province Application Basic Research Plan (20210302124550).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data are contained within the article.

Conflicts of Interest

The authors declare no conflicts of interest.

Appendix A

Table A1. Mapping between ROI names and experiential function networks in AAL template.
Table A1. Mapping between ROI names and experiential function networks in AAL template.
ROIRegionsAbbreviatedRSN
1Precentral_L PreCG.LDMN
2Precentral_R PreCG.RDMN
3Frontal_Sup_L SFGdor.LDMN
4Frontal_Sup_R SFGdor.RDMN
5Frontal_Sup_Orb_L ORBsup.LDMN
6Frontal_Sup_Orb_R ORBsup.RDMN
7Frontal_Mid_L MFG.LAttention
8Frontal_Mid_R MFG.RAttention
9Frontal_Mid_Orb_L ORBmid.LAttention
10Frontal_Mid_Orb_R ORBmid.RAttention
11Frontal_Inf_Oper_L IFGoperc.LAttention
12Frontal_Inf_Oper_R IFGoperc.RAttention
13Frontal_Inf_Tri_L IFGtriang.LAttention
14Frontal_Inf_Tri_R IFGtriang.RAttention
15Frontal_Inf_Orb_L ORBinf.LAttention
16Frontal_Inf_Orb_R ORBinf.RAttention
17Rolandic_Oper_L ROL.LSensorimotor
18Rolandic_Oper_R ROL.RSensorimotor
19Supp_Motor_Area_L SMA.LAttention
20Supp_Motor_Area_R SMA.RSensorimotor
21Olfactory_L OLF.LSubcortical
22Olfactory_R OLF.RDMN
23Frontal_Sup_Medial_L SFGmed.LDMN
24Frontal_Sup_Medial_R SFGmed.RDMN
25Frontal_Mid_Orb_L ORBsupmed.LDMN
26Frontal_Mid_Orb_R ORBsupmed.RDMN
27Rectus_L REC.LDMN
28Rectus_R REC.RDMN
29Insula_L INS.LSensorimotor
30Insula_R INS.RSensorimotor
31Cingulum_Ant_L ACG.LDMN
32Cingulum_Ant_R ACG.RDMN
33Cingulum_Mid_L DCG.LSubcortical
34Cingulum_Mid_R DCG.RSubcortical
35Cingulum_Post_L PCG.LDMN
36Cingulum_Post_R PCG.RDMN
37Hippocampus_L HIP.LSubcortical
38Hippocampus_R HIP.RSubcortical
39ParaHippocampal_L PHG.LSubcortical
40ParaHippocampal_R PHG.RSubcortical
41Amygdala_L AMYG.LSubcortical
42Amygdala_R AMYG.RSubcortical
43Calcarine_L CAL.LVisual
44Calcarine_R CAL.RVisual
45Cuneus_L CUN.LVisual
46Cuneus_R CUN.RVisual
47Lingual_L LING.LVisual
48Lingual_R LING.RVisual
49Occipital_Sup_L SOG.LVisual
50Occipital_Sup_R SOG.RVisual
51Occipital_Mid_L MOG.LVisual
52Occipital_Mid_R MOG.RVisual
53Occipital_Inf_L IOG.LVisual
54Occipital_Inf_R IOG.RVisual
55Fusiform_L FFG.LVisual
56Fusiform_R FFG.RVisual
57Postcentral_L PoCG.LSensorimotor
58Postcentral_R PoCG.RSensorimotor
59Parietal_Sup_L SPG.LSensorimotor
60Parietal_Sup_R SPG.RSensorimotor
61Parietal_Inf_L IPL.LAttention
62Parietal_Inf_R IPL.RAttention
63SupraMarginal_L SMG.LSensorimotor
64SupraMarginal_R SMG.RSensorimotor
65Angular_L ANG.LAttention
66Angular_R ANG.RAttention
67Precuneus_L PCUN.LDMN
68Precuneus_R PCUN.RDMN
69Paracentral_Lobule_L PCL.LSensorimotor
70Paracentral_Lobule_R PCL.RSensorimotor
71Caudate_L CAU.LSubcortical
72Caudate_R CAU.RSubcortical
73Putamen_L PUT.LSubcortical
74Putamen_R PUT.RSubcortical
75Pallidum_L PAL.LSubcortical
76Pallidum_R PAL.RSubcortical
77Thalamus_L THA.LSubcortical
78Thalamus_R THA.RSubcortical
79Heschl_L HES.LSensorimotor
80Heschl_R HES.RSensorimotor
81Temporal_Sup_L STG.LSensorimotor
82Temporal_Sup_R STG.RSensorimotor
83Temporal_Pole_Sup_L TPOsup.LAttention
84Temporal_Pole_Sup_R TPOsup.RSensorimotor
85Temporal_Mid_L MTG.LDMN
86Temporal_Mid_R MTG.RDMN
87Temporal_Pole_Mid_L TPOmid.LSubcortical
88Temporal_Pole_Mid_R TPOmid.RSubcortical
89Temporal_Inf_L ITG.LAttention
90Temporal_Inf_R ITG.RDMN

References

  1. van den Heuvel, M.P.; Fornito, A. Brain networks in schizophrenia. Neuropsychol. Rev. 2014, 24, 32–48. [Google Scholar] [CrossRef] [PubMed]
  2. Lee, W.H.; Doucet, G.E.; Leibu, E.; Frangou, S. Resting-state network connectivity and metastability predict clinical symptoms in schizophrenia. Schizophr. Res. 2018, 201, 208–216. [Google Scholar] [CrossRef] [PubMed]
  3. Fingelkurts, A.A.; Fingelkurts, A.A. Timing in cognition and EEG brain dynamics: Discreteness versus continuity. Cogn. Process 2006, 7, 135–162. [Google Scholar] [CrossRef] [PubMed]
  4. Michel, C.M.; Murray, M.M. Towards the utilization of EEG as a brain imaging tool. Neuroimage 2012, 61, 371–385. [Google Scholar] [CrossRef] [PubMed]
  5. Caria, A.; Sitaram, R.; Birbaumer, N. Real-time fMRI: A tool for local brain regulation. Neuroscientist 2012, 18, 487–501. [Google Scholar] [CrossRef] [PubMed]
  6. Brookes, M.J.; Leggett, J.; Rea, M.; Hill, R.M.; Holmes, N.; Boto, E.; Bowtell, R. Magnetoencephalography with optically pumped magnetometers (OPM-MEG): The next generation of functional neuroimaging. Trends Neurosci. 2022, 45, 621–634. [Google Scholar] [CrossRef] [PubMed]
  7. Mather, M.; Cacioppo, J.T.; Kanwisher, N. Introduction to the Special Section: 20 Years of fMRI-What Has It Done for Understanding Cognition? Perspect. Psychol. Sci. 2013, 8, 41–43. [Google Scholar] [CrossRef] [PubMed]
  8. Chang, C.; Glover, G.H. Time-frequency dynamics of resting-state brain connectivity measured with fMRI. Neuroimage 2010, 50, 81–98. [Google Scholar] [CrossRef]
  9. Hutchison, R.M.; Womelsdorf, T.; Gati, J.S.; Everling, S.; Menon, R.S. Resting-state networks show dynamic functional connectivity in awake humans and anesthetized macaques. Hum. Brain Mapp. 2013, 34, 2154–2177. [Google Scholar] [CrossRef]
  10. Kabbara, A.; El Falou, W.; Khalil, M.; Wendling, F.; Hassan, M. The dynamic functional core network of the human brain at rest. Sci. Rep. 2017, 7, 2936. [Google Scholar] [CrossRef]
  11. Allen, E.A.; Damaraju, E.; Plis, S.M.; Erhardt, E.B.; Eichele, T.; Calhoun, V.D. Tracking whole-brain connectivity dynamics in the resting state. Cereb. Cortex 2014, 24, 663–676. [Google Scholar] [CrossRef] [PubMed]
  12. Li, F.; Peng, W.; Jiang, Y.; Song, L.; Liao, Y.; Yi, C.; Zhang, L.; Si, Y.; Zhang, T.; Wang, F.; et al. The Dynamic Brain Networks of Motor Imagery: Time-Varying Causality Analysis of Scalp EEG. Int. J. Neural Syst. 2019, 29, 1850016. [Google Scholar] [CrossRef] [PubMed]
  13. Cui, X.; Ding, C.; Wei, J.; Xue, J.; Wang, X.; Wang, B.; Xiang, J. Analysis of Dynamic Network Reconfiguration in Adults with Attention-Deficit/Hyperactivity Disorder Based Multilayer Network. Cereb. Cortex 2021, 31, 4945–4957. [Google Scholar] [CrossRef] [PubMed]
  14. Braun, U.; Schäfer, A.; Bassett, D.S.; Rausch, F.; Schweiger, J.I.; Bilek, E.; Erk, S.; Romanczuk-Seiferth, N.; Grimm, O.; Geiger, L.S.; et al. Dynamic brain network reconfiguration as a potential schizophrenia genetic risk mechanism modulated by NMDA receptor function. Proc. Natl. Acad. Sci. USA 2016, 113, 12568–12573. [Google Scholar] [CrossRef] [PubMed]
  15. Zhu, Z.; Wang, H.; Bi, H.; Lv, J.; Zhang, X.; Wang, S.; Zou, L. Dynamic functional connectivity changes of resting-state brain network in attention-deficit/hyperactivity disorder. Behav. Brain Res. 2023, 437, 114121. [Google Scholar] [CrossRef] [PubMed]
  16. Gifford, G.; Crossley, N.; Kempton, M.J.; Morgan, S.; Dazzan, P.; Young, J.; McGuire, P. Resting state fMRI based multilayer network configuration in patients with schizophrenia. Neuroimage Clin. 2020, 25, 102169. [Google Scholar] [CrossRef] [PubMed]
  17. Zhou, Y.; Shu, N.; Liu, Y.; Song, M.; Hao, Y.; Liu, H.; Yu, C.; Liu, Z.; Jiang, T. Altered resting-state functional connectivity and anatomical connectivity of hippocampus in schizophrenia. Schizophr. Res. 2008, 100, 120–132. [Google Scholar] [CrossRef] [PubMed]
  18. Lynall, M.E.; Bassett, D.S.; Kerwin, R.; McKenna, P.J.; Kitzbichler, M.; Muller, U.; Bullmore, E. Functional connectivity and brain networks in schizophrenia. J. Neurosci. 2010, 30, 9477–9487. [Google Scholar] [CrossRef]
  19. van Dellen, E.; Börner, C.; Schutte, M.; van Montfort, S.; Abramovic, L.; Boks, M.P.; Cahn, W.; van Haren, N.; Mandl, R.; Stam, C.J.; et al. Functional brain networks in the schizophrenia spectrum and bipolar disorder with psychosis. NPJ Schizophr. 2020, 6, 22. [Google Scholar] [CrossRef]
  20. Li, J.; Liu, Y.; Wisnowski, J.L.; Leahy, R.M. Identification of overlapping and interacting networks reveals intrinsic spatiotemporal organization of the human brain. Neuroimage 2023, 270, 119944. [Google Scholar] [CrossRef]
  21. Xu, J.; Potenza, M.N.; Calhoun, V.D.; Zhang, R.; Yip, S.W.; Wall, J.T.; Pearlson, G.D.; Worhunsky, P.D.; Garrison, K.A.; Moran, J.M. Large-scale functional network overlap is a general property of brain functional organization: Reconciling inconsistent fMRI findings from general-linear-model-based analyses. Neurosci. Biobehav. Rev. 2016, 71, 83–100. [Google Scholar] [CrossRef] [PubMed]
  22. Cordes, D.; Haughton, V.M.; Arfanakis, K.; Wendt, G.J.; Turski, P.A.; Moritz, C.H.; Quigley, M.A.; Meyerand, M.E. Mapping functionally related regions of brain with functional connectivity MR imaging. AJNR Am. J. Neuroradiol. 2000, 21, 1636–1644. [Google Scholar] [PubMed]
  23. Stam, C.J. Modern network science of neurological disorders. Nat. Rev. Neurosci. 2014, 15, 683–695. [Google Scholar] [CrossRef] [PubMed]
  24. Tzourio-Mazoyer, N.; Landeau, B.; Papathanassiou, D.; Crivello, F.; Etard, O.; Delcroix, N.; Mazoyer, B.; Joliot, M. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage 2002, 15, 273–289. [Google Scholar] [CrossRef] [PubMed]
  25. He, Y.; Wang, J.; Wang, L.; Chen, Z.J.; Yan, C.; Yang, H.; Tang, H.; Zhu, C.; Gong, Q.; Zang, Y.; et al. Uncovering intrinsic modular organization of spontaneous brain activity in humans. PLoS ONE 2009, 4, e5226. [Google Scholar] [CrossRef] [PubMed]
  26. Thompson, W.H.; Fransson, P. A common framework for the problem of deriving estimates of dynamic functional brain connectivity. Neuroimage 2018, 172, 896–902. [Google Scholar] [CrossRef] [PubMed]
  27. Buckner, R.L.; Sepulcre, J.; Talukdar, T.; Krienen, F.M.; Liu, H.; Hedden, T.; Andrews-Hanna, J.R.; Sperling, R.A.; Johnson, K.A. Cortical hubs revealed by intrinsic functional connectivity: Mapping, assessment of stability, and relation to Alzheimer’s disease. J. Neurosci. 2009, 29, 1860–1873. [Google Scholar] [CrossRef]
  28. Fransson, P.; Aden, U.; Blennow, M.; Lagercrantz, H. The functional architecture of the infant brain as revealed by resting-state fMRI. Cereb. Cortex 2011, 21, 145–154. [Google Scholar] [CrossRef]
  29. Fransson, P.; Thompson, W.H. Temporal flow of hubs and connectivity in the human brain. Neuroimage 2020, 223, 117348. [Google Scholar] [CrossRef]
  30. Power, J.D.; Schlaggar, B.L.; Lessov-Schlaggar, C.N.; Petersen, S.E. Evidence for hubs in human functional brain networks. Neuron 2013, 79, 798–813. [Google Scholar] [CrossRef]
  31. Karim, M.R.; Beyan, O.; Zappa, A.; Costa, I.G.; Rebholz-Schuhmann, D.; Cochez, M.; Decker, S. Deep learning-based clustering approaches for bioinformatics. Brief. Bioinform. 2021, 22, 393–415. [Google Scholar] [CrossRef] [PubMed]
  32. van den Heuvel, M.P.; Sporns, O.; Collin, G.; Scheewe, T.; Mandl, R.C.; Cahn, W.; Goñi, J.; Hulshoff Pol, H.E.; Kahn, R.S. Abnormal rich club organization and functional brain dynamics in schizophrenia. JAMA Psychiatry 2013, 70, 783–792. [Google Scholar] [CrossRef] [PubMed]
  33. Dong, D.; Duan, M.; Wang, Y.; Zhang, X.; Jia, X.; Li, Y.; Xin, F.; Yao, D.; Luo, C. Reconfiguration of Dynamic Functional Connectivity in Sensory and Perceptual System in Schizophrenia. Cereb. Cortex 2019, 29, 3577–3589. [Google Scholar] [CrossRef] [PubMed]
  34. Fox, M.D.; Snyder, A.Z.; Vincent, J.L.; Corbetta, M.; Van Essen, D.C.; Raichle, M.E. The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proc. Natl. Acad. Sci. USA 2005, 102, 9673–9678. [Google Scholar] [CrossRef] [PubMed]
  35. de Leeuw, M.; Kahn, R.S.; Zandbelt, B.B.; Widschwendter, C.G.; Vink, M. Working memory and default mode network abnormalities in unaffected siblings of schizophrenia patients. Schizophr. Res. 2013, 150, 555–562. [Google Scholar] [CrossRef] [PubMed]
  36. Whitfield-Gabrieli, S.; Ford, J.M. Default mode network activity and connectivity in psychopathology. Annu. Rev. Clin. Psychol. 2012, 8, 49–76. [Google Scholar] [CrossRef] [PubMed]
  37. Arkin, S.C.; Ruiz-Betancourt, D.; Jamerson, E.C.; Smith, R.T.; Strauss, N.E.; Klim, C.C.; Javitt, D.C.; Patel, G.H. Deficits and compensation: Attentional control cortical networks in schizophrenia. Neuroimage Clin. 2020, 27, 102348. [Google Scholar] [CrossRef]
  38. Wynn, J.K.; Jimenez, A.M.; Roach, B.J.; Korb, A.; Lee, J.; Horan, W.P.; Ford, J.M.; Green, M.F. Impaired target detection in schizophrenia and the ventral attentional network: Findings from a joint event-related potential-functional MRI analysis. Neuroimage Clin. 2015, 9, 95–102. [Google Scholar] [CrossRef]
  39. Kim, S.; Kim, Y.W.; Shim, M.; Jin, M.J.; Im, C.H.; Lee, S.H. Altered Cortical Functional Networks in Patients with Schizophrenia and Bipolar Disorder: A Resting-State Electroencephalographic Study. Front. Psychiatry 2020, 11, 661. [Google Scholar] [CrossRef]
  40. Butler, P.D.; Schechter, I.; Zemon, V.; Schwartz, S.G.; Greenstein, V.C.; Gordon, J.; Schroeder, C.E.; Javitt, D.C. Dysfunction of early-stage visual processing in schizophrenia. Am. J. Psychiatry 2001, 158, 1126–1133. [Google Scholar] [CrossRef]
  41. Schechter, I.; Butler, P.D.; Zemon, V.M.; Revheim, N.; Saperstein, A.M.; Jalbrzikowski, M.; Pasternak, R.; Silipo, G.; Javitt, D.C. Impairments in generation of early-stage transient visual evoked potentials to magno- and parvocellular-selective stimuli in schizophrenia. Clin. Neurophysiol. 2005, 116, 2204–2215. [Google Scholar] [CrossRef] [PubMed]
  42. Butler, P.D.; Zemon, V.; Schechter, I.; Saperstein, A.M.; Hoptman, M.J.; Lim, K.O.; Revheim, N.; Silipo, G.; Javitt, D.C. Early-stage visual processing and cortical amplification deficits in schizophrenia. Arch. Gen. Psychiatry 2005, 62, 495–504. [Google Scholar] [CrossRef] [PubMed]
  43. Mingoia, G.; Wagner, G.; Langbein, K.; Maitra, R.; Smesny, S.; Dietzek, M.; Burmeister, H.P.; Reichenbach, J.R.; Schlösser, R.G.; Gaser, C.; et al. Default mode network activity in schizophrenia studied at resting state using probabilistic ICA. Schizophr. Res. 2012, 138, 143–149. [Google Scholar] [CrossRef] [PubMed]
  44. Hummer, T.A.; Yung, M.G.; Goñi, J.; Conroy, S.K.; Francis, M.M.; Mehdiyoun, N.F.; Breier, A. Functional network connectivity in early-stage schizophrenia. Schizophr. Res. 2020, 218, 107–115. [Google Scholar] [CrossRef] [PubMed]
  45. Harris, K.D.; Mrsic-Flogel, T.D. Cortical connectivity and sensory coding. Nature 2013, 503, 51–58. [Google Scholar] [CrossRef]
  46. Yeo, B.T.; Krienen, F.M.; Chee, M.W.; Buckner, R.L. Estimates of segregation and overlap of functional connectivity networks in the human cerebral cortex. Neuroimage 2014, 88, 212–227. [Google Scholar] [CrossRef]
  47. Northoff, G.; Duncan, N.W. How do abnormalities in the brain’s spontaneous activity translate into symptoms in schizophrenia? From an overview of resting state activity findings to a proposed spatiotemporal psychopathology. Prog. Neurobiol. 2016, 145–146, 26–45. [Google Scholar] [CrossRef]
  48. Xiao, J.; Uddin, L.Q.; Meng, Y.; Li, L.; Gao, L.; Shan, X.; Huang, X.; Liao, W.; Chen, H.; Duan, X. A spatio-temporal decomposition framework for dynamic functional connectivity in the human brain. Neuroimage 2022, 263, 119618. [Google Scholar] [CrossRef]
  49. Northoff, G. Spatiotemporal Psychopathology—A Novel Approach to Brain and Symptoms. Noro Psikiyatr. Ars. 2022, 59, S3–S9. [Google Scholar] [CrossRef]
  50. Hou, C.; Jiang, S.; Liu, M.; Li, H.; Zhang, L.; Duan, M.; Yao, G.; He, H.; Yao, D.; Luo, C. Spatiotemporal dynamics of functional connectivity and association with molecular architecture in schizophrenia. Cereb. Cortex 2023, 33, 9095–9104. [Google Scholar] [CrossRef]
  51. Northoff, G. Is schizophrenia a spatiotemporal disorder of the brain’s resting state? World Psychiatry 2015, 14, 34–35. [Google Scholar] [CrossRef] [PubMed]
  52. Yousefi, B.; Shin, J.; Schumacher, E.H.; Keilholz, S.D. Quasi-periodic patterns of intrinsic brain activity in individuals and their relationship to global signal. Neuroimage 2018, 167, 297–308. [Google Scholar] [CrossRef] [PubMed]
  53. Li, P.; Fan, T.T.; Zhao, R.J.; Han, Y.; Shi, L.; Sun, H.Q.; Chen, S.J.; Shi, J.; Lin, X.; Lu, L. Altered Brain Network Connectivity as a Potential Endophenotype of Schizophrenia. Sci. Rep. 2017, 7, 5483. [Google Scholar] [CrossRef] [PubMed]
  54. Abbas, A.; Bassil, Y.; Keilholz, S. Quasi-periodic patterns of brain activity in individuals with attention-deficit/hyperactivity disorder. Neuroimage Clin. 2019, 21, 101653. [Google Scholar] [CrossRef] [PubMed]
  55. Herlin, B.; Navarro, V.; Dupont, S. The temporal pole: From anatomy to function-A literature appraisal. J. Chem. Neuroanat. 2021, 113, 101925. [Google Scholar] [CrossRef] [PubMed]
  56. Sun, Y.; Collinson, S.L.; Suckling, J.; Sim, K. Dynamic Reorganization of Functional Connectivity Reveals Abnormal Temporal Efficiency in Schizophrenia. Schizophr. Bull. 2019, 45, 659–669. [Google Scholar] [CrossRef] [PubMed]
  57. Tanglay, O.; Young, I.M.; Dadario, N.B.; Briggs, R.G.; Fonseka, R.D.; Dhanaraj, V.; Hormovas, J.; Lin, Y.H.; Sughrue, M.E. Anatomy and white-matter connections of the precuneus. Brain Imaging Behav. 2022, 16, 574–586. [Google Scholar] [CrossRef]
  58. Gao, Y.; Tong, X.; Hu, J.; Huang, H.; Guo, T.; Wang, G.; Li, Y.; Wang, G. Decreased resting-state neural signal in the left angular gyrus as a potential neuroimaging biomarker of schizophrenia: An amplitude of low-frequency fluctuation and support vector machine analysis. Front. Psychiatry 2022, 13, 949512. [Google Scholar] [CrossRef]
  59. Belloy, M.E.; Naeyaert, M.; Abbas, A.; Shah, D.; Vanreusel, V.; van Audekerke, J.; Keilholz, S.D.; Keliris, G.A.; Van der Linden, A.; Verhoye, M. Dynamic resting state fMRI analysis in mice reveals a set of Quasi-Periodic Patterns and illustrates their relationship with the global signal. Neuroimage 2018, 180, 463–484. [Google Scholar] [CrossRef]
  60. Yang, G.J.; Murray, J.D.; Glasser, M.; Pearlson, G.D.; Krystal, J.H.; Schleifer, C.; Repovs, G.; Anticevic, A. Altered Global Signal Topography in Schizophrenia. Cereb. Cortex 2017, 27, 5156–5169. [Google Scholar] [CrossRef]
  61. Karahanoğlu, F.I.; Van De Ville, D. Transient brain activity disentangles fMRI resting-state dynamics in terms of spatially and temporally overlapping networks. Nat. Commun. 2015, 6, 7751. [Google Scholar] [CrossRef] [PubMed]
  62. Kim, G.W.; Kim, Y.H.; Jeong, G.W. Whole brain volume changes and its correlation with clinical symptom severity in patients with schizophrenia: A DARTEL-based VBM study. PLoS ONE 2017, 12, e0177251. [Google Scholar] [CrossRef] [PubMed]
  63. Stevenson, R.J.; Langdon, R.; McGuire, J. Olfactory hallucinations in schizophrenia and schizoaffective disorder: A phenomenological survey. Psychiatry Res. 2011, 185, 321–327. [Google Scholar] [CrossRef] [PubMed]
  64. Masaoka, Y.; Velakoulis, D.; Brewer, W.J.; Cropley, V.L.; Bartholomeusz, C.F.; Yung, A.R.; Nelson, B.; Dwyer, D.; Wannan, C.M.J.; Izumizaki, M.; et al. Impaired olfactory ability associated with larger left hippocampus and rectus volumes at earliest stages of schizophrenia: A sign of neuroinflammation? Psychiatry Res. 2020, 289, 112909. [Google Scholar] [CrossRef] [PubMed]
  65. Seghier, M.L. The angular gyrus: Multiple functions and multiple subdivisions. Neuroscientist 2013, 19, 43–61. [Google Scholar] [CrossRef] [PubMed]
  66. Uddin, L.Q.; Supekar, K.; Amin, H.; Rykhlevskaia, E.; Nguyen, D.A.; Greicius, M.D.; Menon, V. Dissociable connectivity within human angular gyrus and intraparietal sulcus: Evidence from functional and structural connectivity. Cereb. Cortex 2010, 20, 2636–2646. [Google Scholar] [CrossRef]
  67. Popov, T.; Rockstroh, B.; Miller, G.A. Oscillatory connectivity as a mechanism of auditory sensory gating and its disruption in schizophrenia. Psychophysiology 2022, 59, e13770. [Google Scholar] [CrossRef]
  68. Voss, M.; Chambon, V.; Wenke, D.; Kühn, S.; Haggard, P. In and out of control: Brain mechanisms linking fluency of action selection to self-agency in patients with schizophrenia. Brain 2017, 140, 2226–2239. [Google Scholar] [CrossRef]
  69. Leube, D.; Straube, B.; Green, A.; Blümel, I.; Prinz, S.; Schlotterbeck, P.; Kircher, T. A possible brain network for representation of cooperative behavior and its implications for the psychopathology of schizophrenia. Neuropsychobiology 2012, 66, 24–32. [Google Scholar] [CrossRef]
  70. Deschamps, I.; Baum, S.R.; Gracco, V.L. On the role of the supramarginal gyrus in phonological processing and verbal working memory: Evidence from rTMS studies. Neuropsychologia 2014, 53, 39–46. [Google Scholar] [CrossRef]
  71. Barbaro, M.F.; Kramer, D.R.; Nune, G.; Lee, M.B.; Peng, T.; Liu, C.Y.; Kellis, S.; Lee, B. Directional tuning during reach planning in the supramarginal gyrus using local field potentials. J. Clin. Neurosci. 2019, 64, 214–219. [Google Scholar] [CrossRef] [PubMed]
  72. Kong, L.; Herold, C.J.; Cheung, E.F.C.; Chan, R.C.K.; Schröder, J. Neurological Soft Signs and Brain Network Abnormalities in Schizophrenia. Schizophr. Bull. 2020, 46, 562–571. [Google Scholar] [CrossRef] [PubMed]
Brainsci 14 00040 g001
Figure 1. Overview of the schematic diagram of the analysis strategy. (A) Firstly, the standard brain contour present in the AAL template was divided into 90 individual brain regions. Each of these regions represented a node within the network under study. The JC method was then employed to estimate functional connectivity at each time point. (B) Next, the BC method was used to identify candidate hubs within the network that required tracking. The activity of these candidate hubs was observed over time to gain further insight into their interactions with other nodes within the network. Significant differences between SZ and NC were analyzed by examining (C) the spatial overlap, (D) temporal overlap, and hierarchical clustering of hubs.
Figure 1. Overview of the schematic diagram of the analysis strategy. (A) Firstly, the standard brain contour present in the AAL template was divided into 90 individual brain regions. Each of these regions represented a node within the network under study. The JC method was then employed to estimate functional connectivity at each time point. (B) Next, the BC method was used to identify candidate hubs within the network that required tracking. The activity of these candidate hubs was observed over time to gain further insight into their interactions with other nodes within the network. Significant differences between SZ and NC were analyzed by examining (C) the spatial overlap, (D) temporal overlap, and hierarchical clustering of hubs.
Brainsci 14 00040 g001
Brainsci 14 00040 g002
Figure 2. (A) Differences in BC at the whole-brain level. Asterisks represent group differences; ** denotes p < 0.01. (B) Group differences in BC at the RSN level. Asterisks represent the difference between groups; * denotes p < 0.05.
Figure 2. (A) Differences in BC at the whole-brain level. Asterisks represent group differences; ** denotes p < 0.01. (B) Group differences in BC at the RSN level. Asterisks represent the difference between groups; * denotes p < 0.05.
Brainsci 14 00040 g002
Brainsci 14 00040 g003
Figure 3. Distribution of candidate hubs in two groups: (A) NC, (B) SZ.
Figure 3. Distribution of candidate hubs in two groups: (A) NC, (B) SZ.
Brainsci 14 00040 g003
Brainsci 14 00040 g004
Figure 4. (A) Subnetwork distribution of active hubs in NC. (B) Subnetwork distribution of active hubs in SZ patients. The distribution of subnetwork members of active hubs at all time points under the average of all subjects is shown in the form of a pie chart.
Figure 4. (A) Subnetwork distribution of active hubs in NC. (B) Subnetwork distribution of active hubs in SZ patients. The distribution of subnetwork members of active hubs at all time points under the average of all subjects is shown in the form of a pie chart.
Brainsci 14 00040 g004
Brainsci 14 00040 g005
Figure 5. Group differences in spatial overlap. The number (percentage) of spatial co-occurrences for all pairwise comparisons are shown.
Figure 5. Group differences in spatial overlap. The number (percentage) of spatial co-occurrences for all pairwise comparisons are shown.
Brainsci 14 00040 g005
Brainsci 14 00040 g006
Figure 6. Hierarchical clustering based on the spatial pattern of active hubs: (A) NC, (C) SZ. Clusters resulting from hierarchical clustering: (B) NC (D) SZ.
Figure 6. Hierarchical clustering based on the spatial pattern of active hubs: (A) NC, (C) SZ. Clusters resulting from hierarchical clustering: (B) NC (D) SZ.
Brainsci 14 00040 g006
Brainsci 14 00040 g007
Figure 7. Time overlap of active hubs in the two groups of subjects (expressed as a percentage). (A) NC. (B) SZ. Distribution in the right and left brain of large temporal overlap in the top 20%. (C) NC. (D) SZ. The first five significant larger hubs pairs in both groups of subjects. (E) NC. (F) SZ.
Figure 7. Time overlap of active hubs in the two groups of subjects (expressed as a percentage). (A) NC. (B) SZ. Distribution in the right and left brain of large temporal overlap in the top 20%. (C) NC. (D) SZ. The first five significant larger hubs pairs in both groups of subjects. (E) NC. (F) SZ.
Brainsci 14 00040 g007
Table 1. Demographic and clinical characteristics of the participants in this study.
Table 1. Demographic and clinical characteristics of the participants in this study.
CharacteristicSZNCStatistical Test
Number of subjects4349--
Age (years)34.84 ± 8.6033.12 ± 8.22p = 0.331
Sex (male/female)
SAPS		30/13
30.61 ± 20.26		30/19
--		p = 0.112
--
Note: The values are denoted as mean ± standard deviation.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Xiang, J.; Sun, Y.; Wu, X.; Guo, Y.; Xue, J.; Niu, Y.; Cui, X. Abnormal Spatial and Temporal Overlap of Time-Varying Brain Functional Networks in Patients with Schizophrenia. Brain Sci. 2024, 14, 40. https://doi.org/10.3390/brainsci14010040

AMA Style

Xiang J, Sun Y, Wu X, Guo Y, Xue J, Niu Y, Cui X. Abnormal Spatial and Temporal Overlap of Time-Varying Brain Functional Networks in Patients with Schizophrenia. Brain Sciences. 2024; 14(1):40. https://doi.org/10.3390/brainsci14010040

Chicago/Turabian Style

Xiang, Jie, Yumeng Sun, Xubin Wu, Yuxiang Guo, Jiayue Xue, Yan Niu, and Xiaohong Cui. 2024. "Abnormal Spatial and Temporal Overlap of Time-Varying Brain Functional Networks in Patients with Schizophrenia" Brain Sciences 14, no. 1: 40. https://doi.org/10.3390/brainsci14010040

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop