Multiple anatomical brain networks for self-esteem related analysis among young adults

Background: Self-esteem is the individual evaluation of oneself. People with high self-esteem have mental health and can bravely cope with the threats from the environment. With the development of neuroimaging techniques, researches on the cognitive neural mechanisms of self-esteem are increased. Existing methods based on brain morphometry and single-layer brain network cannot characterize the subtle structural differences related to self-esteem. Method: To solve this issue, we proposed multiple anatomical brain network framework based on multi-resolution ROI template to study the cognitive neuroscience mechanism of self-esteem. Multiple anatomical brain network consist of high-resolution ROI features extracted from structural MRI and hierarchal brain network features. For each layer, we calculated the correlation relationship between pairs of ROIs. In order to solve the high-dimensional problem caused by the large amount of multiple network features, combined feature selection methods are adopted to reduce the number of features while retaining discriminative information to the maximum extent. Multi-kernel SVM is employed to combine the two types of features by appropriate weight coecient. Result: Our experimental results show that the proposed method can improve classication accuracy to 97.26% compared with single-layer brain network. Conclusions: This article reveals the cognitive neural mechanism of self-esteem and provides foundation for positive and healthy self-esteem.

Therefore, in this study, we focus on exploring the differences between brain networks of adolescents with different levels of self-esteem.
Brain network aims to study the interaction of various brain regions as a whole, which has an important role to have a deep understanding of brain structures and cognitive neural processes. The anatomical brain network mainly uses the region of interest (ROI) of the brain as the node, and the correlation between brain regions as edge [7][8]. The de nition of ROI is a key step in anatomical brain network analysis. Most existing methods use ROI-based brain network analysis methods to study brain structure and functional connections related to self-esteem. Kelly et al. use the functional near-infrared spectroscopy (fNIR) based cerebral blood ow imaging method to estimate the hemodynamic response function of each ROI, in order to study the brain networks that are activated during the processing of selfesteem related information [9]. Goldin et al. use functional magnetic resonance imaging (fMRI) technology to measure changes in the brain network between the self-esteem group and the selfcon dence group by measuring the BOLD response in the ROI [10]. Chavez et al. conducte a psychophysiological interaction analysis to calculate the correlation between speci c ROIs related to selfesteem [11]. Although a variety of neuroimaging methods can be used to explore the cognitive mechanism of the brain structural magnetic resonance imaging (sMRI) is widely used in the analysis of brain anatomical networks due to its high resolution of brain soft tissue imaging [12]. Studies based on sMR show that self-esteem involves multiple networks related to self-reference processing, autobiographical memory, and social cognition, including default mode networks and social cognition networks [13]. In addition, self-esteem shows the brain network mechanism dominated by bilateral brain and mainly controlled by right brain [14]. Although the above researches have initially revealed the brain network representation of self-esteem, it only used single-network that cannot fully identify the subtle differences in network connectivity caused by self-esteem.
In order to better describe the network representation of self-esteem, enhanced feature representation method is required to better examine function connectivity related to self-esteem. In recent years, machine learning techniques become a research hotspot in the eld of brain network analysis due to its ability to learn laws from data and predict unknown data [8]. Brain network analysis can help us fully understand the cognitive psychological activity of self-esteem. However, there are few studies using machine learning methods to construct self-esteem related brain networks, especially for the construction of multi-layer brain networks. In this article, we propose a novel multiple anatomical brain network construction method based on multi-resolution ROIs, which can better describe the correlation between small brain regions and large brain functional areas simultaneously.

Classi cation performance
Various indexes can be used to evaluate the classi cation performance of the proposed method (Fig. 1). The evaluation indicators include accuracy, sensitivity, speci city, area under the receiver operating characteristic curve, F score, balanced accuracy, Youden's index are listed in Table 1. The research results show that the multi-layer brain network features have the highest classi cation accuracy of 97.26%, and the AUC is also greater than other feature types. This indicates that the multiple brain network features have advantages in characterizing structural differences at the global level. In addition, the higher speci city and sensitivity also show that the multiple brain network features have better recognition capabilities in exploring the subtle differences in brain structure caused by self-esteem.

Weight coe cient
The role of the weight coe cient is to determine the proportion of the two types of features in the multikernel classi er (Fig. 2). Appropriate weight coe cient help the classi er performance the best. A smaller weight coe cient indicates that the contribution of the fourth-layer ne ROI features is lower, while the contribution of the hierarchical brain network features is higher. Through experiments, we can nd the most suitable weight coe cient in the range of 0-1.
The weight coe cient has an important in uence on the performance of the classi er. It is proved that the weight coe cient can make the classi er perform well in the relatively large range from 0.05 to 0.35, which can decline the di culty of determining the ratio of the two features, which re ects the robustness of our proposed method. The best results are obtained at 0.05. At this time, the hierarchical brain network features contributed more to the classi cation than the high-resolution ROI features in the bottommost layer. This is because the hierarchical brain network can fully express the differences in brain structure between the two groups.

Top discriminative features
We use the proposed method to select the most discriminative ROI features (Fig. 3). These ROIs include occipital lobe (superior and middle occipital gyrus, cuneus), frontal lobe (supplementary motor area, middle frontal gyrus), temporal lobe (middle temporal gyrus), parental lobe (precuneus, angular gyrus), limbic lobe (posterior cingulate gyrus), and central region (precentral gyrus). The experimental results also show that differences in brain structure related to self-esteem are mainly in WM and cortical thickness ( Table 2). The top 15 network features selected from all four layers (Table 3). The most discriminative hierarchical network features are mainly distributed in limbic lobe and parental lobe (Fig. 4).

Discussion
We studied multiple anatomical brain network related to self-esteem. Our results have demonstrated that the proposed method is superior to the single network method. The multiple networks enhance the representation of the speci c brain structure related to self-esteem, thereby providing an effective and novel method to detect self-esteem related biomarkers.

Improvement of the proposed method
It is di cult to fully understand the functional organization of the brain using only a single network framework since the brain is a complex system. In this study, we construct a multiple anatomical brain network in multi-resolution ROIs to improve the classi cation performance. Compared with the singlenetwork based method, multiple networks enhance the classi cation performance by using supplementary information from different networks. Compared with the best results obtained using a single network, our proposed multiple anatomical network method can improve the classi cation accuracy by 8.95% (Table 1).

Analysis of discriminative features
The discriminative ROI features discovered by our method are distributed in multiple regions of the brain. Because few current studies employ automatic classi cation method to study the brain structure of selfesteem, we only compare brain regions found through our machine learning method with existing morphological based studies. Compared with previous studies, our results showed consistency in departmental brain regions, including precuneus [4], precentral gyrus [15], middle frontal gyrus [16], cuneus [4], posterior cingulate [17], angular [18]. This indicates the effectiveness of our classi cation method in revealing brain regions related to self-esteem. In addition to these consistent regions, we also found that the middle occipital, superior occipital, and supplementary motor are related to self-esteem. These brain regions have not been reported in previous studies.
The discriminative network features are mainly located on frontal, parental and limbic lobe. After a comprehensive analysis of existing research on neuropsychological mechanisms related to self-esteem, we found that the frontal region is an important part of the neural basis related to self-esteem. The frontal lobe is responsible for self-evaluation, self-regulation, and emotion management. Individuals with low self-esteem have a stronger emotional response to social evaluations, while high self-esteem individuals show stronger self-positivity in the process of self-evaluation. These ndings indicate that frontal lobe plays an important role in generating positive self-information.

Conclusions
In this study, we proposed a multiple anatomical brain network based on sMRI among adolescent.
Compared with the single network structure, the features extracted from the proposed methods can improve the self-esteem related network representation. Both high-resolution ROI features and the hierarchical network features contribute to the improvement of the classi er. The optimal hierarchical network features provide us a new perspective to inspect the discriminative regions of self-esteem. The results of cross-validation experiments also prove the effectiveness of our method. In subsequent research, other brain cognitive activity research and brain disease diagnosis can be carried out by the multiple anatomical brain networks.

Subjects
The structural sMRI data used in our study were acquired from the Soochow University, which is composed of 68 adolescents. The study was approved by the Ethics Committee of the Third A liated Hospital of Soochow University. Written informed consents was obtained from all subjects. Each subject was interviewed by a psychologist to rule out any mental or neurological diseases. No subjects had received stimulant or hypnotics before. All participants' vision was normal or corrected to normal, and they were right-handed. After the test, each participant will receive a small gift or nancial reward. All subjects are required to perform Rosenberg Self-esteem Scale (RSES) test. The RSES is originally developed by Rosenberg in 1965 to assess the overall feelings of adolescents about self-worth and selfacceptance. It is the most used self-esteem measurement tool in the psychology community [15]. We ranked the RSES test scores from highest to lowest, and then divided them into two groups: high selfesteem group and low self-esteem group. Table 4 provides detailed information of all participants. We use an automatic pipeline for sMRI image processing. Firstly, we adjusted the image orientation (axial, coronal, and sagittal) to match the template image, and performed offset eld correction to remove the gray-scale unevenness of the image [19]. Secondly, the brain was extracted by removing the skull and cerebellum [20]. Thirdly, gray matter (GM), white matter (WM) and cerebrospinal uid (CSF) were segmented from the background [21]. Fourth, the segmented image was registered to the template labeled with the Automated Anatomical Labeling (AAL) template [22]. Fifth, in order to calculate the morphological features based on the cortex, the middle layer of the cerebral cortex was constructed [23]. After the whole processing, the morphological measurements of GM volume, WM volume, CSF volume, cortical thickness, and cortical surface area of each ROI were obtained for each subject. It should be noted that we removed 12 subcortical ROIs from AAL template considering that the cerebral cortex contains more neurons.

Classi cation framework
The framework of the proposed classi cation algorithm based on multi-resolution ROI brain network is shown in Fig. 5, mainly including multiple anatomical network construction, feature selection, and classi cation. Multi-resolution ROI based multiple anatomical brain network were constructed based on morphological features (volume of different brain tissue, cortical thickness, and cortical surface area). Feature selection can reduce the dimensionality of high-dimensional brain network features, only retaining the features that can maximize the speci city of the subjects. The optimal feature subset can be trained by the classi er as neuroimaging markers representing different self-esteem levels.

Construction of multiple anatomical networks
Through the above image processing steps, GM volume, WM volume, CSF volume, cortical thickness, and cortical surface area of each ROI can be obtained from the MRI image of each subject. In order to reduce individual differences, standardization was performed, dividing the measured value of each ROI by the total intracranial volume, mean cortical thickness, and whole cerebral cortical surface area of the subject.
Therefore, we used normalized volume and cortical features to provide a more appropriate representation. More objective measurements can be received by such processing. In order to improve the performance of the classi er, we propose a four-layer hierarchical network framework in this paper. We used brain templates with different ROI resolution in each layer to construct brain network nodes and edges.
Speci cally, the bottommost template containing 78 ROIs is de ned as , the remaining three layers are de ned as , where . A larger value indicates a higher-resolution ROI, which is located in the brain network layer closer to the bottom of the hierarchy. By merging small brain regions into large brain functional areas, the number of ROIs are reduced. In the layer , there are 36 ROIs by dividing the whole brain into lateral, medial and inferior surfaces. In the layer , 14 ROIs are de ned ree ng to the anatomical brain structure of central, frontal, parietal, occipital, temporal, limbic, and insula lobe. The speci c de nition rules of these ROIs can be found in Table 5. It is worth noting that in the rst layer , we study the brain as a whole.

Feature selection
In order to reduce the feature dimension and lter out the most discriminative features, we adopted a combined feature selection method. First, we use the statistical t-test method for preliminary selection of features with the signi cant p value less than the threshold (p <0.05). Then, the redundant features are removed using the minimum redundancy and maximum correlation (mRMR) method, and only the features that can express the difference between groups in the minimum number are retained [24]. After the above two lter-based feature selections, the machine learning recursive feature elimination (SVM-RFE) method [25] is used to further reduce the feature dimension. After completing the entire feature selection steps, the optimal feature subset is obtained.
Classi cation using multi-kernel SVM There are two types of features in the multiple brain network, one is the high-resolution ROI features in the fourth layer, and the other is the brain network features corresponding to different layers. Multi-kernel machine learning method can integrate these two types of features into a single classi er. Firstly, a Gaussian Radial Basis Function (RBF) kernel function is used to construct a kernel matrix for each type of feature. Secondly, the two kernel matrices are integrated into the multi-kernel matrix through appropriate weight coe cients [25]. Comparing the results of using linear kernel function and using RBF function (non-linear), we discover that the RBF kernel can signi cantly improve the classi cation performance. Therefore, we choose the RBF kernel function to construct the multi-kernel classi er. Finally, the optimal features subset can be obtained.

Cross-validation
The nested cross-validation method has been applied in our previous research. In the inner loop, the training set are used to determine the parameters of the classi er. In the outer loop, the testing set is used to evaluate the generalization ability of the classi er. It should be noted that at the beginning of the experiment, the entire data set was randomly divided into two parts, one for training and the other one for testing. The training set and testing set can be exchanged throughout the veri cation process, while the processing steps remain unchanged.
Abbreviations ROI: the region of interest; fNIR: the functional near-infrared spectroscopy; fMRI: functional magnetic resonance imaging; sMRI: structural magnetic resonance imaging; RSES: Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
The study is approved by the Ethics Committee of the Third A liated Hospital of Soochow University.

Consent for publication
All subjects gave written informed consent in accordance with the Declaration of Helsinki.

Figure 3
The most discriminating ROI features projected onto the cortical surface.
Page 17/18 Correlative matrix. (a) high self-esteem group, (b) low self-esteem group. (c) differences between the two groups.