Multilayer Network Analysis Research Articles

Multilayer network analysis for the interconnectedness between financial sectors and new energy companies in China

Multilayer network analysis for the interconnectedness between financial sectors and new energy companies in China

The impact of social media discourse on financial performance of e-commerce companies listed on Borsa Istanbul

This research investigates how changes in public discourse on social media (particularly Twitter, now X) influence financial performance of companies listed on the e-commerce XTCRT index of Borsa Istanbul amid the COVID-19 pandemic. Utilising a unique 4-layer network structure, the study integrates daily closing prices, highest and lowest values, and trading volume data of these companies with Twitter discussions on six categories: culture, economy, health, politics, technology, world affairs. By using advanced computational methods such as multilayer network analysis, long short-term memory architecture, nonlinear autoregressive distributed lag analysis, we find a high responsiveness of the e-commerce index to social media content, particularly in discussions related to health, technology, world affairs. The study underscores that social media discussions significantly shape index composition, resulting in noticeable shifts in market dynamics, including fluctuations in stock prices, trading volumes, and market sentiment. Our study advocates for a more advanced approach to market analysis that incorporates the dynamics of online public discussions.

Taiwan’s scientific collaboration under the New Southbound Policy: Trends, effects, mechanisms, and implications

My dissertation explores the evolving trends and patterns in Taiwan’s scientific collaboration under the New Southbound Policy (NSP), with a focus on research co-authorship with NSP target countries across Southeast and South Asia, as well as Australia and New Zealand. Launched in 2016, the NSP represents Taiwan’s strategic effort to reintegrate into the Indo-Pacific region amidst shifting geopolitical realities. This study investigates the underexplored dynamics of scientific collaboration and the seldom-tested policy effects in a non-Western, non-colonial context, exemplifying science diplomacy in action. Specifically, this research addresses three key questions: (1) What are the long-term trends in research collaboration with NSP target countries from the mid-2000s to 2021? (2) How do co-authorship patterns and dynamics manifest in cross-specialization profiles and authorship orders among these countries? (3) What impact has the NSP had on the volume of co-authored research? Methodologically, this study combines bibliometric, network, and modeling approaches. It employs established bibliometric indicators for measuring attractiveness and preferences among partnered countries, multilayer network analysis, and piecewise models of change. The primary data sources are drawn from Clarivate’s Web of Science, accessed in both aggregated forms via the Observatoire des sciences et des technologies and as raw JSON data retrieved through the Web of Science Expanded API. By analyzing the patterns, effects, and mechanisms of Taiwan’s scientific collaboration with NSP target countries, this dissertation contributes to broadening theories and methodologies of international scientific collaboration. It also sheds light on South–South collaboration and offers policy implications for nations navigating similar geopolitical challenges.

A multilayer network analysis of Alzheimer's disease pathogenesis: Roles for p-tau, synaptic peptides, and physical activity.

In the aging brain, cognitive abilities emerge from the coordination of complex pathways arising from a balance between protective lifestyle and environmental factors and accumulation of neuropathologies. As part of the Rush Memory and Aging Project (n=440), we measured accelerometer-based actigraphy, cognitive performance, and after brain autopsy, selected reaction monitoring mass spectrometry. Multilevel network analysis was used to examine the relationships among the molecular machinery of vesicular neurotransmission, Alzheimer's disease (AD) neuropathology, cognition, and late-life physical activity. Synaptic peptides involved in neuronal secretory function were the most influential contributors to the multilayer network, reflecting the complex interdependencies among AD pathology, synaptic processes, and late-life cognition. Older adults with lower physical activity evidenced stronger adverse relationships among phosphorylated tau peptides, markers of synaptic integrity, and tangle pathology. Network-based approaches simultaneously model interdependent biological processes and advance understanding of the role of physical activity in age-associated cognitive impairment. Network-based approaches simultaneously model interdependent biological processes. Secretory synaptic peptides were influential contributors to the multilayer network. Older adults with lower physical activity had adverse relationships among pathology. There was interdependence among phosphorylated tau, synaptic integrity, and tangles. Network methods elucidate the role of physical activity in cognitive impairment.

Using multilayer network analysis to detect the collaborative knowledge construction characteristics among learner groups with low, medium, and high levels of cognitive engagement

Collaborative knowledge construction (CKC) is advanced by group members' cognitive engagement across three levels: individual-level knowledge processing and proposing, the peer-level social interactions and knowledge exchanges between two group members, and the group-level coordination and construction of knowledge across multiple group members. The interconnected and transformative relationship among three levels is an essential factor during the CKC process. Individual viewpoints can trigger peer feedback, which can be further refined by other group members; the group-level consensus can provide foundations for subsequent individual- or peer-level cognitive engagement. The group-level coordination of knowledge advancement involves the processing, synthesis, and reflection of knowledge from multiple members to reach a group consensus. However, previous analytical approaches have faced challenges in modeling distinct types of interconnections across the individual, peer, and group levels during the CKC process. To address this gap, this research employed multilayer network analysis (MNA) to model and quantify the multi-level and interconnected characteristics of CKC in a series of face-to-face, computer-supported CKC activities in China's higher education. First, small groups were categorized into the high, medium, and low levels of cognitive engagement groups. Second, multilayer networks were constructed for each category, where students' use of cognitive strategies were set as nodes, the interconnections between cognitive strategies were set as edges, and three levels (i.e., individual, peer, and group levels) were set as layers. The node-, layer-, and network-level metrics were calculated to quantify the overall characteristics of the networks, the interconnected characteristics within a level and across the three levels. The MNA results revealed that, compared to lower level of cognitive engagement groups, groups with higher cognitive engagement demonstrated (1) more influential moderate and deep cognitive strategies at group and peer levels; (2) interconnections between cognitive strategies with higher diversity, connectivity, and more balanced distribution across three levels, but relatively lower information exchanging efficiency; and (3) communities with more interconnections related to the moderate and deep cognitive strategies at peer and group levels. Based on the empirical findings, this research proposed pedagogical implications for CKC practices and analytical implications for using MNA to improve the understanding of computer-supported collaborative learning mechanisms.

Multilayer network analysis reveals instability of brain dynamics in untreated first-episode schizophrenia.

Although aberrant static functional brain network activity has been reported in schizophrenia, little is known about how the dynamics of neural function are altered in first-episode schizophrenia and are modulated by antipsychotic treatment. The baseline resting-state functional magnetic resonance imaging data were acquired from 122 first-episode drug-naïve schizophrenia patients and 128 healthy controls (HCs), and 44 patients were rescanned after 1-year of antipsychotic treatment. Multilayer network analysis was applied to calculate the network switching rates between brain states. Compared to HCs, schizophrenia patients at baseline showed significantly increased network switching rates. This effect was observed mainly in the sensorimotor (SMN) and dorsal attention networks (DAN), and in temporal and parietal regions at the nodal level. Switching rates were reduced after 1-year of antipsychotic treatment at the global level and in DAN. Switching rates at baseline at the global level and in the inferior parietal lobule were correlated with the treatment-related reduction of negative symptoms. These findings suggest that instability of functional network activity plays an important role in the pathophysiology of acute psychosis in early-stage schizophrenia. The normalization of network stability after antipsychotic medication suggests that this effect may represent a systems-level mechanism for their therapeutic efficacy.

Analyzing critical core components’ technology opportunities based on multilayer networks from a lifecycle perspective: A case study of offshore wind turbine foundation

Among renewable energy sources, offshore wind power has emerged as an important trend due to its more stable and powerful wind resources. Advances in the fundamental technology of offshore wind turbines are crucial for ensuring the long-term feasibility and reliability of wind power projects. Existing technology opportunity analysis (TOA) studies based on link prediction network analysis seldom consider the technical characteristics of the target component from different perspectives, such as lifecycle. In this study, we propose a framework to analyze technology opportunities for offshore wind turbine foundations. The method consists of four steps: 1) Collect patents related to the target component. 2) Use BERT to automatically layer patents into corresponding lifecycle stages. 3) Build a multilayer network centered on the target component. 4) Use link prediction to recommend technology opportunities throughout the lifecycle and at each stage of the target component. Using offshore wind power foundations as an example, the effectiveness of the method is verified. Our findings indicate that improvements in structural arrangements of foundations and anti-corrosion coatings are future technology opportunities for offshore wind power foundations. Multilayer network analysis based on link prediction from a lifecycle perspective provides a valuable reference for future TOA research and practice. Additionally, it offers enterprises insights into future technology opportunities at various lifecycle stages, presenting a new perspective for optimizing R&D resource allocation and market positioning based on these results.

Multilayer network analysis of idiosyncratic volatility connectedness: Evidence from China

This paper proposes multilayer networks, including lagged, contemporaneous, and long-run networks, to examine the idiosyncratic connectedness among Chinese financial institutions. We explore the topology of multilayer networks through static and dynamic analysis to capture the information transmission mechanism of idiosyncratic risks. Moreover, we investigate the drivers influencing idiosyncratic connectedness across financial institutions. We find that idiosyncratic risk contagion among financial institutions is stronger in the long-run network. At the same time, we note that the three networks contain different connection structures, which means that the information transmission mechanism of idiosyncratic risks is heterogeneous in multilayer networks. In addition, when the financial system is under stress, we observe that the contemporaneous effect of idiosyncratic connectedness rises rapidly. Institutional level analysis shows that diversified financial institutions play active roles in the lagged network, securities institutions are transmitters of contemporaneous information spillovers, and banking institutions are the main drivers of the long-run network. Finally, our study shows that size is the main factor driving idiosyncratic risk spillovers among financial institutions.

Determination of electron screening potential of 6Li(p, [formula omitted])3He reaction using Multi Layer Perceptron based neural network

Understanding nuclear reactions between light-charged nuclei in the sub-Coulomb energy region is crucial for several astrophysical processes. Accurate determination of the reaction cross-section within the astrophysically important Gamow range is challenging due to electron screening. Various methods, including polynomial fits, R-Matrix, and the indirect Trojan Horse Method (THM), have estimated electron screening energies that exceed the adiabatic limit. This study aims to derive the bare astrophysical S-factor for the reaction 6Li(p,α)3He and to extract electron screening energies using Multi-Layer Perceptron-based Artificial Neural Network (ANN) analysis. Experimental S-factors for 6Li(p,α)3He, obtained from the literature, are reanalyzed with the ANN algorithm to determine the energy-dependent S-factor. The bare astrophysical S-factor is calculated from data above 60 keV, where electron screening is negligible. The electron screening potential is then derived by comparing the shielded S-factor with the bare S-factor. The ANN-based analysis yields an electron screening potential of 220 eV, suggesting that ANN could be a viable tool for estimating electron screening potentials in light nuclei reactions.

Exploring Social Networks: An Analysis of Intra-organizational Networks

This research investigates intra-organizational social networks by employing network science methods, aiming to provide actionable insights for both managers and employees. Using datasets from the Colorado Index of Complex Networks (ICON), this study examines four intra-organizational networks from a consulting firm and a research team, focusing on advice requests and skill awareness. The analysis includes network topology, degree distribution, community detection, correlation between communities and node attributes, maximal clique detection, and multilayer network analysis. The findings reveal that intra-organizational networks are intricate and significantly impact organizational efficiency and individual career development. Key insights include the importance of mid-level employees in network connectivity, the existence of "bad edges" where advice is sought from less competent peers, and the influence of regional and locational factors on communication patterns. The study emphasizes the need to enhance knowledge sharing and inter-region communications. For employees, understanding the structure of these networks can guide them to more effectively leverage their social connections for career advancement. This interdisciplinary effort bridges gaps in the existing literature and showcases the potential of integrating network science methods into social science research.

Sex differences in the flexibility of dynamic network reconfiguration of autism spectrum disorder based on multilayer network.

Dynamic network reconfiguration alterations in the autism spectrum disorder (ASD) brain have been frequently reported. However, since the prevalence of ASD in males is approximately 3.8 times higher than that in females, and previous studies of dynamic network reconfiguration of ASD have predominantly used male samples, it is unclear whether sex differences exist in dynamic network reconfiguration in ASD. This study used resting-state functional magnetic resonance imaging data from the Autism Brain Imaging Data Exchange database, which included balanced samples of 64 males and 64 females with ASD, along with 64 demographically-matched typically developing control (TC) males and 64 TC females. The multilayer network analysis was used to explore the flexibility of dynamic network reconfiguration. The two-way analysis of variance was further performed to examine the sex-related changes in ASD in flexibility of dynamic network reconfiguration. A diagnosis-by-sex interaction effect was identified in the cingulo-opercular network (CON), central executive network (CEN), salience network (SN), and subcortical network (SUB). Compared with TC females, females with ASD showed lower flexibility in CON, CEN, SN, and SUB. The flexibility of CEN and SUB in males with ASD was higher than that in females with ASD. In addition, the flexibility of CON, CEN, SN, and SUB predicted the severity of social communication impairments and stereotyped behaviors and restricted interests only in females with ASD. These findings highlight significant sex differences in the flexibility of dynamic network reconfiguration in ASD and emphasize the importance of further study of sex differences in future ASD research.

Multilayer network analysis in patients with juvenile myoclonic epilepsy.

We conducted a multilayer network analysis in patients with juvenile myoclonic epilepsy (JME) and healthy controls, to investigate the gray matter layer using a morphometric similarity network and analyze the white matter layer using structural connectivity. We enrolled 42 patients with newly diagnosed JME and 53 healthy controls. Brain magnetic resonance imaging (MRI) using a three-tesla MRI scanner, including T1-weighted imaging and diffusion tensor imaging (DTI) were performed. We created a gray matter layer matrix with a morphometric similarity network using T1-weighted imaging, and a white matter layer matrix with structural connectivity using the DTI. Subsequently, we performed a multilayer network analysis by applying graph theory. There were significant differences in network at the global level in the multilayer network analysis between the groups. The average multiplex participation of patients with JME was lower than that of healthy controls (0.858 vs. 0.878, p = 0.007). In addition, several regions showed significant differences in multiplex participation at the nodal level in the multilayer network analysis. Multiplex participation in the right entorhinal cortex was lower, whereas multiplex participation in the right supramarginal gyrus was higher at the nodal level in the multilayer network analysis of patients with JME compared to healthy controls. We demonstrated differences in network at the global and nodal levels in the multilayer network analysis between patients with JME and healthy controls. These features may be associated with the pathophysiology of JME and could help us understand the complex brain network in patients with JME.

Exploring animal behaviour multilayer networks in immersive environments - a conceptual framework.

Animal behaviour is often modelled as networks, where, for example, the nodes are individuals of a group and the edges represent behaviour within this group. Different types of behaviours or behavioural categories are then modelled as different yet connected networks which form a multilayer network. Recent developments show the potential and benefit of multilayer networks for animal behaviour research as well as the potential benefit of stereoscopic 3D immersive environments for the interactive visualisation, exploration and analysis of animal behaviour multilayer networks. However, so far animal behaviour research is mainly supported by libraries or software on 2D desktops. Here, we explore the domain-specific requirements for (stereoscopic) 3D environments. Based on those requirements, we provide a proof of concept to visualise, explore and analyse animal behaviour multilayer networks in immersive environments.

Mitigating risks in central line-associated bloodstream infection: a comprehensive insight into critical care nurses’ knowledge, attitudes, barriers, and compliance

BackgroundCentral venous catheter-related bloodstream infections (CLABSIs) are a significant concern in intensive care units (ICUs) as they lead to increased morbidity, mortality, and healthcare costs. Fortunately, these infections are largely preventable through strict adherence to CLABSI prevention guidelines. Nurses play a critical role in preventing CLABSIs.AimThis study aimed to investigate factors affecting critical care nurses' knowledge, attitudes, and perceived barriers related to implementing CLABSI prevention guidelines, and to predict factors influencing compliance with these guidelines.MethodsThis cross-sectional study was conducted from April to May 30, 2023, with a convenience sample of 470 critical care nurses from ICUs across eight hospitals in Sana’a, Yemen. Data were collected using an observational checklist and self-administered questionnaire. Descriptive statistics, Independent Student’s t-test, one-way ANOVA, Pearson’s correlation coefficient, multiple linear regression, and multilayer perceptron neural networks were performed.ResultsCritical care nurses exhibited low knowledge of CLABSI prevention guidelines, with compliance reaching an acceptable level. Despite the higher perceived barriers, the nurses demonstrated a positive attitude. Nurses with greater knowledge and positive attitudes displayed higher compliance levels. However, perceived barriers were negatively associated with knowledge and compliance. Notably, multilayer neural network analysis identified knowledge and perceived barriers as the strongest predictors of nurses' compliance.ConclusionThe current findings emphasize the need for multifaceted strategies to implement the CLABSI prevention guidelines. These strategies should address knowledge gaps, support positive attitudes, and address practical barriers faced by nurses to ensure successful implementation of CLABSI prevention.

Extreme risk contagions among fossil energy companies in China: Insights from a multilayer dynamic network analysis

Since the severity and frequency of extreme events have seriously threatened the stability of fossil energy markets, this paper conducts multilayer dynamic network analysis to demonstrate the evolution of the entire market structure during extreme events. Based on the data of 64 Chinese energy firms ranging from March 2018 to February 2023, this paper examines the sectoral discrepancies, and identifies systemically important companies during different crises. The results indicate that the risk transmission mechanisms among fossil energy markets are different induced by the two crises, namely the COVID-19 pandemic and the Russia–Ukraine conflict. Specifically, the COVID-19 has rendered greater influence on the oil and gas industry, while the Russia–Ukraine conflict has brought instability to the coal industry. Additionally, the heterogeneous effects of these events on individual energy firms are identified, and the results suggest that both small and large firms have played prominent roles in risk transmission.

Climate policy uncertainty and the Chinese sectoral stock market: A multilayer network analysis

Climate change is the most important challenge currently faced by humans, which affects their health and financial stability. This study proposes a multilayer network that includes return, volatility, and risk layers to investigate the risk connectedness between climate policy uncertainty and the sectoral stock market in China. We explore the static and dynamic topological features of multilayer networks from the perspectives of the system and market levels, respectively. The results indicate that (i) a crisis increases the intensity of risk spillovers; (ii) an obvious nonsynchronous effect was found among the layers during financial turmoil; and (iii) industrials, materials, information technology, and consumer discretionary act more often as net risk transmitters during the sample period, whereas climate policy uncertainty, telecommunication services, and financials are more frequently net risk receivers. The results of this study have important implications for regulators, policymakers, and investors in alleviating climate risk and developing hedging and investment strategies.

EEG-based driving intuition and collision anticipation using joint temporal-frequency multi-layer dynamic brain network.

Intuition plays a crucial role in human driving decision-making, and this rapid and unconscious cognitive process is essential for improving traffic safety. We used the first proposed multi-layer network analysis method, "Joint Temporal-Frequency Multi-layer Dynamic Brain Network" (JTF-MDBN), to study the EEG data from the initial and advanced phases of driving intuition training in the theta, alpha, and beta bands. Additionally, we conducted a comparative study between these two phases using multi-layer metrics as well as local and global metrics of single layers. The results show that brain region activity is more stable in the advanced phase of intuition training compared to the initial phase. Particularly in the alart state task, the JTF-MDBN demonstrated stronger connection strength. Multi-layer network analysis indicates that modularity is significantly higher for the non-alert state task than the alert state task in the alpha and beta bands. In the W4 time window (1 second before a collision), we identified significant features that can differentiate situations where a car collision is imminent from those where no collision occurs. Single-layer network analysis also revealed statistical differences in node strength and local efficiency for some EEG channels in the alpha and beta bands during the W4 and W5 time windows. Using these biomarkers to predict vehicle collision risk, the classification accuracy of a linear kernel SVM reached up to 87.5%, demonstrating the feasibility of predicting driving collisions through brain network biomarkers. These findings are important for the study of human intuition and the development of brain-computer interface-based intelligent driving hazard perception assistance systems.

Exploring the Complex Relationship between Social Network Addiction and Mental Health in College Students: A Study Based on Multi-layer Network Analysis

With the development of science and technology, various types of online social media have developed rapidly, attracting a large number of users, but excessive use of social network addiction phenomenon, affecting the life and mental health. For college students, social network addiction is not uncommon and affects their mental health, so it is very important to understand the relationship between social network addiction and mental health and to intervene, so this study aims to explore the relationship between social network addiction and mental health. In this study, 1141 college students completed the Social Networking Addiction Scale and the College Mental Health Scale through an online survey. Social network addiction and mental health were then included as nodes and network analysis was performed using R software. It was found that, in the node network graph, there was a link between the well-being experience dimension of mental health and the mood change dimension of social network addiction, and there was a link between the courage to challenge dimension of mental health and the online social interaction preference dimension of social network addiction. In the total sample and the male sample, "I feel that my life is boring" in the mental health scale is the node with the highest intensity, while "I can't concentrate on my study" in the mental health scale has the highest node intensity in the female sample. From the above results, it is concluded that there is a relationship between some dimensions of social network addiction and some dimensions of mental health among college students. Individuals' boredom affects their mental health and social network addiction, and in the case of female college students, their commitment to study also affects their mental health and social network addiction.

Machine learning and integrative multi-omics network analysis for survival prediction in acute myeloid leukemia

BackgroundAcute myeloid leukemia (AML) is the most common malignant myeloid disorder in adults and the fifth most common malignancy in children, necessitating advanced technologies for outcome prediction. MethodThis study aims to enhance prognostic capabilities in AML by integrating multi-omics data, especially gene expression and methylation, through network-based feature selection methodologies. By employing artificial intelligence and network analysis, we are exploring different methods to build a machine learning model for predicting AML patient survival. We evaluate the effectiveness of combining omics data, identify the most informative method for network integration and compare the performance with standard feature selection methods. ResultsOur findings demonstrate that integrating gene expression and methylation data significantly improves prediction accuracy compared to single omics data. Among network integration methods, our study identifies the best approach that improves informative feature selection for predicting patient outcomes in AML. Comparative analyses demonstrate the superior performance of the proposed network-based methods over standard techniques. ConclusionsThis research presents an innovative and robust methodology for building a survival prediction model tailored to AML patients. By leveraging multilayer network analysis for feature selection, our approach contributes to improving the understanding and prognostic capabilities in AML and laying the foundation for more effective personalized therapeutic interventions in the future.

Withdrawn: The altered dynamic community structure for adaptive adjustment in stroke patients with multidomain cognitive impairments: A multilayer network analysis

BackgroundInternal capsule strokes can lead to multidomain cognitive impairments, and while the brain showcases adaptability post-stroke, the interplay between this adaptability and multidomain cognitive function remains elusive. In addition, although most studies indicate that adaptive changes after stroke manifest as brain functional network abnormalities, the changes in dynamic topological features, especially dynamic community structure, required for adaptive adjustment in stroke patients are not yet clear. MethodsResting-state fMRI data were acquired from patients with infarct in left-sided (n = 46, CI_L) and right-sided (n = 45, CI_R) internal capsule and 81 healthy controls (HC). Dynamic network matrices were generated using a sliding-window approach. These matrices were partitioned into communities utilizing a multilayer community detection algorithm. The network switching rate were employed as a key metric to characterize the differences of dynamic community structure between stroke and HC groups, and the correlations between these significant intergroup differences and cognitive performance scores was calculated. ResultsCompared to HC, CI_L patients exhibited an increase in network switching rates across all networks in the whole brain. However, these increases were largely unbeneficial for cognitive recovery; CI_R group show a more limited range of increased network switching rates and a more limited negative correlations with cognitive performance. ConclusionsThis study elucidates that following a stroke, the brain continuously adapts its dynamic community structure in an effort to address multidomain cognitive impairments, but this adaptation is often ineffective, unorganized, and disordered. Moreover, these findings elucidate the lateralized effects of stroke lesions on dynamic reorganization.