Abnormal Functional and Structural networks in Rectal Cancer with Depressive risk: A Graph Theory Analysis

Background: Surgery and chemotherapy can cause depressive risk in patients with rectal cancer (RC). However, few comprehensive studies are conducted on RC patients associated alterations induced by emotional disorders in the topological organization of structural and functional networks. Methods: Resting-state functional MRI and Diffusion tensor imaging data were collected from 36 RC patients with surgery and chemotherapy and 32 healthy controls (HC). Functional network (FN) was constructed from extracting average time courses for 246 regions of interest (ROI) and structural network (SN) was established by deterministic tractography. Graph theoretical analysis was used to calculate small-worldness property, clustering coefficients, shortest path length and network efficiency. Additionally, we assess network resilient on FN and SN. Results: Abnormal small-worldness property of FN and SN were found in RC patients. The FN and SN exhibited increased local efficiency and global efficiency respectively in RC patients.The increased nodal efficiency in RC patients were mainly found in the frontal lobe, parietal lobe and limbic lobe for FN and SN, while the decreased nodal efficiency were distributed in subcortical nuclei, parietal lobe and limbic lobe only for SN. In network resilient analysis, the RC patients showed less resilient to targeted or random node deletion in both networks compared with HC. Moreover, FN is more robust than SN for all participants. Conclusions: This study revealed that topological organizations of the FN and SN may be disrupted in RC patients. Brain network reorganization is a compensation mechanism to alleviate the depressive risk in RC patients after surgery and chemotherapy.

that a considerable number of cancer patients with surgery and chemotherapy had morphological variation and functional abnormalities in brain, such as decreased hippocampal volume [8], lower white matter volume [9], as well as memory difficulties [2] and cognitive deficit [10]. Therefore, research on FN and SN in this study could bring new insights into the neurophysiological mechanisms in RC patients, and through the analysis of brain images, the emotional distress of RC patients could be diagnosed and treated, thereby improving the quality of life of patients [11].
Combining multi-modal data can reveal hidden relationships among different data, unifying different findings in brain imaging [12]. Therefore, multi-modal imaging is a prominent method in cognitive neuroscience research. Graph-based functional and structural brain connectivity analysis is a new method, which provides evidences for the complexity of the brain by modeling the interactions between different brain regions [13]. Previous studies have consistently shown that brain functional networks are organized in a small-worldness property with local specialization and high capacity for global information transfer [14,15]. Bruno, Hosseini [2] reported significantly reduced shortest path length and small-worldness property in the breast cancer group. Several findings point that the functional network of cancer patients loses its ability to support various cognitive functions following chemotherapy [16,17]. Observational studies found that alterations in brain structural network had an adverse impact on the cognition of cancer survivors [18,19]. Although there are many studies on brain cognitive impairment in cancer patients, little is known about FN and SN abnormalities in RC patients. We used multimodal neuroimaging to investigate the alterations in functional and structural connectivity for getting a deep understanding of the brain cognitive dysfunction in cancer patients.
In view of the poor understanding of psychological disorders and cognitive impairment of rectal cancer survivors in existing studies, it is essential to study the depressive risk and related factors in rectal cancer patients. Therefore, the present study investigated abnormalities in FN and SN using graph theory analysis in RC patients with surgery and chemotherapy characterized by depressive risk compared with healthy controls (HC). We hypothesized that RC patients would show altered smallworldness property and topological architecture in the FN and SN due to the effects of depressive risk.
We sought to expand our understanding of the resilience of the brain network in RC patients. The study also explored the potential association between the significant alterations in network properties of RC patients and severity of depression symptoms.

Participants
36 RC patients were recruited from the Gansu Provincial Hospital, while the 32 age and gender matched healthy control participants were recruited through newspaper advertisements. They were recruited from July 2017 to May 2019. All participants were diagnosed according to DSM-IV criteria by two experienced psychiatrist. They have executed the evaluation of 17-item Hamilton Rating Scale for Depression (HAMD-17). We divided the severity of depression in cancer patients into three categories: HAMD score 1-7 (17 participants), score 8-17 (14 participants), score>17 (5 participants). All participants were given written informed consent when image scanning.

Data Processing
All rs-fMRI data were preprocessed using the Statistical Parametric Mapping (SPM8: http://www.fil.ion.ucl.ac.uk/spm) and Data Processing Assistant Resting-State fMRI (DPARSFA; http://www.restfmri.net ) [20]. The specific preprocessed steps were as follows: (1) the first 10 volumes of the functional images were removed; (2) slice timing, head motion correction, and realignment were performed; (3) all subjects were excluded if their head motion was great than 2.0 mm maximum displacement in any of the x, y or z directions was great than 2° [21]; (4) the leaved rs-fMRI data of 32 HC and 36 RC patients were spatially normalized to Montreal Neurological Institute (MNI) space by applying the parameters of structural image normalization (resampling voxel size of 3mm×3mm×3mm); (5) smoothing with an 8 mm Gaussian kernel of 8 mm full width at half maximum (FWHM) [22]; (6) nuisance covariates regression including 24 head motion parameters, averaged global, white matter signals and cerebrospinal fluid [23]; (7) removing linear trends;; (8) temporal band-pass filtering (0.01-0.08 Hz) was performed to reduce low-frequency drift and high-frequency physiological noise [24].

Construction of Brain Networks
The nodes of the network were demarcated according to the Human Brainnetome Atlas (246 Atlas) with 210 cortical and 36 subcortical subregions [26]. In each subject, 246 Atlas were used to construct brain networks for further graph theory analysis. Functional brain network was constructed using GRETNA (www.nitrc.org/projects/gretna/), which is a toolbox for analyzing brain connections. For each subject, a 246*246 temporal correlation matrix was constructed using the mean time series of each region by Pearson's correlation coefficient. In order to convert the data into z-values for normal distribution, Fisher's z-transform was performed to each matrix. A FA weighted symmetric matrix (246*246) was constructed for each participant by deterministic tractography as the structural network for following research. Each matrix represented the white matter network of the cerebral cortex, and each row or column in network represented the brain region of 246 atlas. For each participant, FN and SN were used for further graph analysis.

Threshold calculation
In order to construct an undirected binary network and make the generated graph metrics stable, it is necessary to be thresholded for the weight of the brain networks. There is no fixed method to determine the threshold in current research. Therefore, in FN, we used sparsity (26% 50%) with a step of 1% [27] to divide the network threshold. Then, we calculated the topological properties of FN in a series of thresholds range. In SN, we use FA (0.2 0.42) with a step of 0.2 as the threshold of the network according to previous study [28]. Because unfully connected networks and highly connected networks have an impact on small-worldness property [29], we use these sparsities to determine the network density reasonably.

Whole brain network organization
Graph theoretical analyses of the FN and SN in RC patients and HC were calculated with routines from the GRETNA toolbox. The network topological properties at the global levels were calculated, including (1) properties that suggest network segregation of brain, such as the normalized clustering coefficient (γ), the local efficiency ; (2) properties that indicate network integration of the brain, such as the normalized shortest path length(λ), the global efficiency ; (3) small-worldness (δ) property which evaluates the balance of segregation and integration.
The nodal efficiency ( ) measures the ability of a particular node to propagate information with all other nodes in the network. It is considered as the inverse of the harmonic mean of the minimum path length between an index node and all other nodes in the network.

Network Resilience Analysis
Network resilience refers to the ability to withstand perturbations or failures in the network, which is usually related to the stability of complex networks [30,31]. In FN and SN, we used random or targeted attacks with fixed sparsity or FA values to evaluate the network resilience, so as to ensure that all anatomical regions were involved in the network, thus minimizing the number of false-positive paths [30]. In targeted attack analysis, the betweenness value of each node in the network were calculated and sorted in descending order. We deleted the nodes in the network in order of betweenness value and calculated the global efficiency of each network after attack [32]. In random attacks analysis, we deleted the nodes of network randomly and calculated the global efficiency of each network after attack.

Statistical Analysis
The demographic and clinical characteristics of the RC patients and HC were analyzed by Chi square test and two-sample t-tests using SPSS 21. We performed statistical comparisons of topological measures between the two groups using non-parametric permutation tests with 5000 iterations for each sparsity and FA value [33]. For the , the non-parametric permutation tests was repeated at a fixed sparsity (sp=26%) and a FA value (FA=0.42). FDR correction was conducted for all these results.
Besides, we used Pearson correlation analyses to explore the correlations in RC patients between nodes with significant difference in and the severity of depression (HAMD score).
There were significant difference in the HAMD (p<0.001) score between the two groups.

Global Topology of Functional and Structural networks
For the rs-fMRI datasets, compared with the HC, RC patients showed a higher shortest path length (λ) (Figure1B, sparsity=26%) and unchanged clustering coefficient (γ) Figure 1A , which resulted in abnormal small-worldness( ) ( Figure 1C). Additionally, RC patients showed increased E loc ( Figure 1D, sparsity =26%) and unchanged E glob ( Figure 1E). Original P value and the FDR-corrected P value of topological measures for each sparsity in Supplementary materials.

Regional efficiency analysis
Compared with the HC, RC patients showed that significantly decreased nodal efficiency was only in FN. There were several regions including bilateral basal ganglia, right parahippocampal gyrus, bilateral thalamus, right precuneus, and right lateral occipital cortex. Meanwhile, the increased nodal efficiency was mainly in frontal lobe (orbital gyrus), basal ganglia, left inferior frontal gyrus, left amygdala, bilateral cingulate gyrus, left inferior parietal lobule, and right precentral gyrus in FN and SN for RC patients (p<0.05, after 5000 permutation test, FDR test)( Table 2, Table 3; Figure 3).

The comparison of network resilience
With the targeted and random attack, a significantly decreased decline of the global efficiency was found in FN and SN (Figure 4). In both networks, the global efficiency of RC patients decreased faster over a wide percentage of removal, which reflected that the networks of RC patients were more fragile. In all subjects, the resilience of structural network is weaker than that of functional network under the same threshold.

Discussion
In this study, we explored different topological organizations of FN and SN in RC patients and HC. The findings pointed RC patients displayed altered small-worldness property and global topological organization compared with HC. Moreover, there were regions with significant abnormal being mainly distributed in frontal region, subcortical regions and central region in RC patients. In addition, RC patients showed vulnerable network resilience in both networks, and FN would be more stable than SN across participants.

Network Properties
Although the global and regional brain network properties in breast cancer and lung cancer patients are reported in neuroimaging research using fMRI [34,35] and DTI [19], rectal carcinoma is still little.
Compared with HC, the functional networks of RC patients displayed a higher shortest path length (λ) and decreased small-worldness( ), reflecting reduced global integration and disrupted organization balance [2,19]. Our results also revealed increased local efficiency in RC patients. It is a measure of local information transmission among adjacent nodes and therefore an indication of network segregation [36]. Previous studies demonstrated reduced local efficiency, a common measure of the brain network's response to computational attack, associated with breast cancer patients [19,37].
Due to brain structural damage, decreased local efficiency would affect the fault tolerant ability of brain network. More detail, the result of weakening network fault tolerance is that if a node in the brain is damaged, the connection between previously linked nodes would be greatly affected [38].
Therefore, reduced local efficiency is a risk factor for RC patients. Recently, researchers use graph theory to analyze complex brain functional networks after chemotherapy. It has been proved that chemotherapy-related cognitive deficits were associated with abnormal topological alterations of brain functional and structural network [39][40][41]. In this study, increased shortest path length and decreased local efficiency in RC patients with surgery and chemotherapy could be seen as a brain compensation mechanism, which included changing the global pathway and adjusting regional activity to preserve a seesaw-like balance of the brain network.
RC patients showed increased clustering coefficients (γ), small-worldness( ) and global efficiency in SN ( Figure 2). Abnormal small-worldness property of SN indicated that the local specialization and global integration of brain in RC patients were disrupted, where the SN tended to be more randomized [42]. Global efficiency is the inverse of the average shortest path between nodes. When nodes could interact directly, the efficiency is high [19]. Therefore, global efficiency is an indicator of network function integration and parallel information processing capability [38]. The present results of abnormal network properties reflected the undesired topological organization in SN, which exhibited that the deficits of emotional and cognitive processing in RC patients might result from network damages. Besides, the increased network properties of SN in RC patients might suggest that local nerve fibers reconstructed in response to the abnormalities in brain functional network. The compensatory response of the SN is activated for maintaining brain functional integrity to compensate the cognitive impairment caused by chemotherapy to RC patients [43]. Aforementioned evidences illuminated that cognitive deficit related to RC patients may act via disrupted coordination between global and regional networks.

Regional nodal parameters
To explore the functional and structural characteristics of the human brain more accurately and quantitatively, our study employed a new standard brain atlas, containing 246 brain regions. This atlas would allow brain network analysis to use predefined nodes in an informed manner [44].
Therefore, more detailed division of brain regions provide better help in multi-modal data analysis.
We observed decreased only in FN of RC patients. The significantly changed regions were located in bilateral basal ganglia, bilateral thalamus, right parahippocampal gyrus, right precuneus, and right lateral occipital cortex. The basal ganglia is not only related to motor control, but also related to the cognitive and limbic functions [44]. Moreover, basal ganglia is the collection of subcortical nuclei surrounding the thalamus [45]. Abnormal activation of basal ganglia/thalamus was found in the depressive studies [46,47], suggesting that abnormalities in these brain regions may lead to abnormal emotional processing mechanisms. Prior studies reported that parahippocampal gyrus and precuneus were associated with memory function, so alterations in these regions might affect revealed that the activation of frontal and parietal lobes increased during the speech working memory task 1 month after chemotherapy. Compared with controls, the cancer group showed significantly greater activation in right precentral gyrus, right cingulate gyrus [17]. Moreover, in the SN, the nodal efficiency were only increased. We speculated that after surgery and chemotherapy, the node efficiency of SN showed more obvious activation in order to maintain robustness of overall network at the expense of other network property, such as integration. These results improved the understanding of chemotherapy-induced cognitive impairment in RC patients from the perspective of brain node efficiency.
As shown in Figure 5, the RC patients showed a positive relationship between HAMD and decreased nodal efficiency in mPMtha.R of FN, as well as a positive relationship between HAMD and increased nodal efficiency in LAmyg.L of SN. The correlation between the changed node efficiency and HAMD score may indicate impaired cognitive control combined with abnormal affective processing in RC patients [58]. A prior study suggested that regions sensitive to negative emotions were hyperactive in processing negative information [59], it was not surprising to find a significant positive correlation between increased nodal efficiency and HAMD in the amygdala. Moreover, the positive relationship between HAMD and decreased nodal efficiency revealed that abnormal activation of FN in RC patients might cause cognitive impairments and depressed mood [60]. Therefore, we speculated that alterations in brain network properties assist us to study the depressive risk in RC patients after chemotherapy and surgery.

The difference of network resilience
In both networks, a key finding of significantly decreased resilience to targeted and random attack was found (Figure 4). Being more effective than other network properties to measure network integration performance, global efficiency of the FN and SN were utilized to explore network resilience quantitatively [32]. In the present study, both networks of RC patients were more vulnerable and SN is less resilient than FN, which were consistent with our previous research results [61]. This finding enhanced the conclusion that lower brain resilience was associated with progressive deterioration of cognitive impairment in breast cancer survivors [19]. Similar results were investigated in other neurological diseases such as major depressive disorders [31] and temporal lobe epilepsy [30]. A previous study showed that the degree distribution of brain network followed the exponentially truncated power law [62]. This exponentially truncated power law distribution may be helpful in resisting the targeted attack of the hubs, meaning the brain networks of two groups were almost constant when deletion rate was low [63]. The deletion ratios reaching 50%, the decline rate of global efficiency in networks began to exhibit obvious differences. Exploring the resilient of networks actually simulated the process of cognitive decline in all participants. In detail, as the important nodes were deleted, the functional and structural integrity of brain networks were impaired. Additionally, the FN was more resilient than the SN in present study, which were similar with these findings in previous studies [28,64]. A prior study discovered that there was commonly a functional connectivity between regions that have no direct structural connectivity, implying that functional network was a more stable system in brain network [65]. Therefore, functional networks were more robust to node removal. Our results may provide a new direction for studying cognitive impairment in RC patients after surgery and chemotherapy.

Conclusions
The present study uncovers the effects of depression symptoms on brain functional and structural network in RC patients with surgery and chemotherapy through multimodal brain connectivity analysis. RC patients show the abnormal small-worldness property and network topological organization in FN and SN. The alterations in nodal parameter are mainly observed in the limbic and parietal lobes as well as the subcortical nuclei in RC patients. The RC patients demonstrate significant cognitive impairment compared with HC and this impairment may be associated with lower network attack tolerance. The discovery of functional and structural networks is critical for understanding the neurobiological mechanism associated with depressive symptoms in RC patients with surgery and chemotherapy.

Limitations
The cross-sectional study limited the ability of studying the causal relationship between alterations in brain network and depressive symptoms of RC patients. The statistical power is restricted by small sample size to some extent. Finally, this study lacks the joint analysis for multimodal data. It is very meaningful to use different modal data for fusion research. Declarations research protocol has been approved by the Gansu Provincial Hospital Ethics Committee.

Consent for publication
Not applicable.       Regions with significant differences in nodal efficiency between RC patients and HC.
Nonparametric permutation tests were applied to nodal efficiency of all 246 cortical regions (p < 0.05 5000 permutation test, FDR correction). A represented FN, and B represented SN.
Red is for increased nodal efficiency in RC patients group, while blue is for decreased nodal efficiency in RC patients group. L=left; R=right.