Information Flow In Networks Research Articles

Altered dynamic network interactions in children with ASD during face recognition revealed by time-varying EEG networks.

Although the electrophysiological event-related potential in face processing (e.g. N170) is widely accepted as a face-sensitivity biomarker that is deficient in children with autism spectrum disorders, the time-varying brain networks during face recognition are still awaiting further investigation. To explore the social deficits in autism spectrum disorder, especially the time-varying brain networks during face recognition, the current study analyzed the N170, cortical activity, and time-varying networks under 3 tasks (face-upright, face-inverted, and house-upright) in autism spectrum disorder and typically developing children. The results revealed a smaller N170 amplitude in autism spectrum disorder compared with typically developing, along with decreased cortical activity mainly in occipitotemporal areas. Concerning the time-varying networks, the atypically stronger information flow and brain network connections across frontal, parietal, and temporal regions in autism spectrum disorder were reported, which reveals greater effort was exerted by autism spectrum disorder to obtain comparable performance to the typically developing children, although the amplitude of N170 was still smaller than that of the typically developing children. Different brain activation states and interaction patterns of brain regions during face processing were discovered between autism spectrum disorder and typically developing. These findings shed light on the face-processing mechanisms in children with autism spectrum disorder and provide new insight for understanding the social dysfunction of autism spectrum disorder.

Investigating the Spatial Heterogeneity and Influencing Factors of Urban Multi-Dimensional Network Using Multi-Source Big Data in Hangzhou Metropolitan Circle, Eastern China

Exploring the spatial heterogeneity of urban multi-dimensional networks and influencing factors are of great significance for the integrated development of metropolitan circle. This study took Hangzhou metropolitan circle as an example, using multi-source geospatial big data to obtain urban population, transportation, goods, capital, and information flow information among sub-cities. Then, spatial visualization analysis, social network analysis, and geographical detector were applied to analyze the differences in spatial structure of multiple urban networks and influencing factors in Hangzhou metropolitan circle, respectively. The results showed that (1) the network connections of population, traffic, goods, and capital flows transcended geographical proximity except that of information flow, and population and traffic flow networks were found to be more flattened in Hangzhou metropolitan circle than in other urban networks; (2) the comprehensive urban network of Hangzhou metropolitan circle was imbalanced across sub-cities, presenting hierarchical and unipolar characteristics; and (3) the influence of traffic distance on the network spatial structure of Hangzhou metropolitan was stronger than the geographical distance, and the interactions between traffic distance and socioeconomic factors would further enhance the regional differentiation of the network spatial structure. This study could provide scientific reference for constructing a coordinated and integrated development pattern in a metropolitan circle.

Decomposing and Tracing Mutual Information by Quantifying Reachable Decision Regions.

The idea of a partial information decomposition (PID) gained significant attention for attributing the components of mutual information from multiple variables about a target to being unique, redundant/shared or synergetic. Since the original measure for this analysis was criticized, several alternatives have been proposed but have failed to satisfy the desired axioms, an inclusion-exclusion principle or have resulted in negative partial information components. For constructing a measure, we interpret the achievable type I/II error pairs for predicting each state of a target variable (reachable decision regions) as notions of pointwise uncertainty. For this representation of uncertainty, we construct a distributive lattice with mutual information as consistent valuation and obtain an algebra for the constructed measure. The resulting definition satisfies the original axioms, an inclusion-exclusion principle and provides a non-negative decomposition for an arbitrary number of variables. We demonstrate practical applications of this approach by tracing the flow of information through Markov chains. This can be used to model and analyze the flow of information in communication networks or data processing systems.

Revealing stock market risk from information flow based on transfer entropy: The case of Chinese A-shares

According to the efficient-market hypothesis, asset prices can reflect all available information in an efficient stock market. This paper focuses on the transmission of information flow between different equities during a financial crisis to reveal the potential risk spreading in the stock market. The information flow and corresponding time-varying stock networks (i.e., information flow networks and industry networks) are constructed using the transfer entropy with the intra-day data of the Chinese A-Share stock market around the 2008 subprime mortgage crisis. The empirical results show that the information flow can reflect the market condition and the financial crisis stage well. The information flow is more apparent when it falls than when it rises. Similarly, it is more apparent in the crash period than in the downturn period. Furthermore, the evolution of information flow networks, markedly shows that several crucial topological parameters, indicates the crash period. We identify four key stocks, through which the whole stock market tends to exchange market information. Finally, we transform the information flow networks into coarse-grained industry networks, and find that the stock market is more likely to be driven by a master or two industries. Through the core industries, the entire stocks are more like to exchange market information, resulting in the increasing system risk in the crash period. Our work provides an interesting insight for regulators to manage market risks and for market participants to avoid troubles.

Trauma apps and the making of the ‘smart’ refugee

A market for mental health apps, designed to help millions of refugees manage symptoms of Post Traumatic Syndrome Disorder and other mental health issues, has proliferated since the outbreak of the so-called refugee crisis in 2015. These bite-size, on-the-go, mindfulness-based apps have emerged at the intersection of new investment models, state-of-the-art AI and surveillance and border control regimes. Conceived of as a more cost-effective approach to refugee mental health care, mental health apps are part of a larger endeavour to create the 'smart' refugee. Self-monitoring, agile, entrepreneurial and resilient in the face of adversity, the smart refugee is expected to emerge as a node in a network of information flow, constantly connected to digital technology, at once receiving and providing real-time data. Biometric and data markets, some of the fastest growing in the world, have already been eagerly collecting refugee fingerprints, iris scans, facial images and other genomic information. To add to this arsenal of data, the new apps are harvesting, storing and selling what I call the mental prints of refugee trauma, turning the human experience of loss, grief and suffering into quantifiable and marketable commodities.

Research on non-smooth signal identification algorithm for electronic communication in a two-way relay network

Abstract To better allow accurate identification of non-smooth signals in two-way relay networks. In this paper, an algorithm based on the two-way relay network system model, the RLS algorithm, is proposed. The algorithm introduces the recursive least squares principle algorithm to the basic cost function of nonsmooth signal identification, optimizes the signal identification model by exponential weighting, and uses the time-varying variance matrix and the optimal identification data separation matrix to make real-time estimates of the new cost function for obtaining nonsmooth signals. And a new index, the PI index, is introduced to define whether the non-stationary signal is stable and reliable. Through two sets of comparison tests, we can learn that the RLS algorithm based on the two-way relay network system is better than other algorithms in the identification of non-stationary signals. It shows that the RLS algorithm can better identify and manage the information flow in a two-way relay network, make the non-stationary signal more stable, and further improve the convergence and convergence speed of the non-stationary signal to provide a better information interaction experience for both sides of electronic communication.

Frequency-domain physical constrained neural network for nonlinear system dynamic prediction

The current use of physical constraints as prior knowledge of neural network models is a new frontier for artificial intelligence to solve complex nonlinear system predictions. Aiming at the problems of limited feature expression and low time-domain computational efficiency of existing physical-constrained neural networks, a physical informed neural network (PINN) modeling approach in frequency domain for nonlinear system dynamic prediction is proposed. First, the frequency-domain physics-constrained neural network (FPNN) framework is constructed, which mainly consists of CliqueNet and Fourier spectral method (FSM). Then, the CliqueNet with convolution feature cyclic structure for high dimensional sequential data is designed. There are forward and backward connections between any two layers of the network to improve the cross-level information flow of the deep network. Finally, FSM is utilized to inform the neural network of physical states by learning specific initial conditions without training data. Simulation results show that FPNN can effectively achieve Kuramoto–Sivashinsky complex system state prediction, with higher computational efficiency and faster learning efficiency compared with SOTA approaches.

Time-varying characteristics of information flow networks in the Chinese market: An analysis based on sector indices

This study analyses the dynamics of the information flow between sector indices in the Chinese market. Calculations show that the effective transfer entropy matrix is time-varying and stable over most periods, and that a few critical events strongly affect the information flow dynamics. We analyse the dynamics using IS (influence strength)-analysis and find that abnormal IS values were accompanied by high market volatility and major events. In particular, we find that the dominant information source changes drastically over time in the sequence of information flow networks, suggesting that the dominant sector is volatile.

Fairness of Information Flow in Social Networks

Social networks form a major parts of people’s lives, and individuals often make important life decisions based on information that spreads through these networks. For this reason, it is important to know whether individuals from different protected groups have equal access to information flowing through a network. In this article, we define the Information Unfairness (IUF) metric, which quantifies inequality in access to information across protected groups. We then introduce MinIUF , an algorithm for reducing inequalities in information flow by adding edges to the network. Finally, we provide an in-depth analysis of information flow with respect to an attribute of interest, such as gender, across different types of networks to evaluate whether the structure of these networks allows groups to equally access information flowing in the network. Moreover, we investigate the causes of unfairness in such networks and how it can be improved.

Time-Lagged Multidimensional Pattern Connectivity (TL-MDPC): An EEG/MEG pattern transformation based functional connectivity metric

Functional and effective connectivity methods are essential to study the complex information flow in brain networks underlying human cognition. Only recently have connectivity methods begun to emerge that make use of the full multidimensional information contained in patterns of brain activation, rather than unidimensional summary measures of these patterns. To date, these methods have mostly been applied to fMRI data, and no method allows vertex-to-vertex transformations with the temporal specificity of EEG/MEG data. Here, we introduce time-lagged multidimensional pattern connectivity (TL-MDPC) as a novel bivariate functional connectivity metric for EEG/MEG research. TL-MDPC estimates the vertex-to-vertex transformations among multiple brain regions and across different latency ranges. It determines how well patterns in ROI X at time point tx can linearly predict patterns of ROI Y at time point ty. In the present study, we use simulations to demonstrate TL-MDPC's increased sensitivity to multidimensional effects compared to a unidimensional approach across realistic choices of number of trials and signal-to-noise ratios. We applied TL-MDPC, as well as its unidimensional counterpart, to an existing dataset varying the depth of semantic processing of visually presented words by contrasting a semantic decision and a lexical decision task. TL-MDPC detected significant effects beginning very early on, and showed stronger task modulations than the unidimensional approach, suggesting that it is capable of capturing more information. With TL-MDPC only, we observed rich connectivity between core semantic representation (left and right anterior temporal lobes) and semantic control (inferior frontal gyrus and posterior temporal cortex) areas with greater semantic demands. TL-MDPC is a promising approach to identify multidimensional connectivity patterns, typically missed by unidimensional approaches.

Effective Connectivity Evaluation of Resting-State Brain Networks in Alzheimer's Disease, Amnestic Mild Cognitive Impairment, and Normal Aging: An Exploratory Study.

(1) Background: Alzheimer's disease (AD) is a neurodegenerative disease with a high prevalence. Despite the cognitive tests to diagnose AD, there are pitfalls in early diagnosis. Brain deposition of pathological markers of AD can affect the direction and intensity of the signaling. The study of effective connectivity allows the evaluation of intensity flow and signaling pathways in functional regions, even in the early stage, known as amnestic mild cognitive impairment (aMCI). (2) Methods: 16 aMCI, 13 AD, and 14 normal subjects were scanned using resting-state fMRI and T1-weighted protocols. After data pre-processing, the signal of the predefined nodes was extracted, and spectral dynamic causal modeling analysis (spDCM) was constructed. Afterward, the mean and standard deviation of the Jacobin matrix of each subject describing effective connectivity was calculated and compared. (3) Results: The maps of effective connectivity in the brain networks of the three groups were different, and the direction and strength of the causal effect with the progression of the disease showed substantial changes. (4) Conclusions: Impaired information flow in the resting-state networks of the aMCI and AD groups was found versus normal groups. Effective connectivity can serve as a potential marker of Alzheimer's pathophysiology, even in the early stages of the disease.

Exploring the Temporal Patterns of Dynamic Information Flow during Attention Network Test (ANT)

The attentional processes are conceptualized as a system of anatomical brain areas involving three specialized networks of alerting, orienting and executive control, each of which has been proven to have a relation with specified time-frequency oscillations through electrophysiological techniques. Nevertheless, at present, it is still unclear how the idea of these three independent attention networks is reflected in the specific short-time topology propagation of the brain, assembled with complexity and precision. In this study, we investigated the temporal patterns of dynamic information flow in each attention network via electroencephalograph (EEG)-based analysis. A modified version of the attention network test (ANT) with an EEG recording was adopted to probe the dynamic topology propagation in the three attention networks. First, the event-related potentials (ERP) analysis was used to extract sub-stage networks corresponding to the role of each attention network. Then, the dynamic network model of each attention network was constructed by post hoc test between conditions followed by the short-time-windows fitting model and brain network construction. We found that the alerting involved long-range interaction among the prefrontal cortex and posterior cortex of brain. The orienting elicited more sparse information flow after the target onset in the frequency band 1-30 Hz, and the executive control contained complex top-down control originating from the frontal cortex of the brain. Moreover, the switch of the activated regions in the associated time courses was elicited in attention networks contributing to diverse processing stages, which further extends our knowledge of the mechanism of attention networks.

Consensus reaching for social network group decision making with ELICIT information: A perspective from the complex network

Consensus reaching is essential in group decision-making (GDM) since it can mitigate conflicts between expert opinions and promotes the further implementation of decision-making results. Meanwhile, interaction between experts commonly occurs within social networks and in practical GDM problems. Therefore, it neecessary to consider the trust relationship between experts and utilize it to facilitate the consensus-reaching process (CRP). However, most existing social network-based GDM studies mainly use local measures (e.g., degree centrality) to determine the importance of experts, which cannot reflect their actual influence on a global topological structure. To address this issue, we propose a novel consensus-reaching strategy from the perspective of complex network analysis. First, the Extended Comparative Linguistic Expressions with Symbolic Translation (ELICIT) is adopted to flexibly facilitate the expression of experts’ uncertain evaluations. The hybrid centrality is then defined to determine the influence of experts in the social network by considering both node importance and edge weight. Since experts with greater influence have stronger information propagation capabilities, hybrid centrality is utilized to guide the CRP, which can better reflect information flows in the social network. Additionally, the BWM-CRITIC weighting method is developed to reflect the significance and relationship among criteria. Finally, we verify the effectiveness and superiority of the proposed method by means of a case study on a sustainable supplier selection problem.

Electroencephalographic Effective Connectivity Analysis of the Neural Networks during Gesture and Speech Production Planning in Young Adults.

Gestures and speech, as linked communicative expressions, form an integrated system. Previous functional magnetic resonance imaging studies have suggested that neural networks for gesture and spoken word production share similar brain regions consisting of fronto-temporo-parietal brain regions. However, information flow within the neural network may dynamically change during the planning of two communicative expressions and also differ between them. To investigate dynamic information flow in the neural network during the planning of gesture and spoken word generation in this study, participants were presented with spatial images and were required to plan the generation of gestures or spoken words to represent the same spatial situations. The evoked potentials in response to spatial images were recorded to analyze the effective connectivity within the neural network. An independent component analysis of the evoked potentials indicated 12 clusters of independent components, the dipoles of which were located in the bilateral fronto-temporo-parietal brain regions and on the medial wall of the frontal and parietal lobes. Comparison of effective connectivity indicated that information flow from the right middle cingulate gyrus (MCG) to the left supplementary motor area (SMA) and from the left SMA to the left precentral area increased during gesture planning compared with that of word planning. Furthermore, information flow from the right MCG to the left superior frontal gyrus also increased during gesture planning compared with that of word planning. These results suggest that information flow to the brain regions for hand praxis is more strongly activated during gesture planning than during word planning.

Identification of vital chemical information via visualization of graph neural networks.

Qualitative or quantitative prediction models of structure-activity relationships based on graph neural networks (GNNs) are prevalent in drug discovery applications and commonly have excellently predictive power. However, the network information flows of GNNs are highly complex and accompanied by poor interpretability. Unfortunately, there are relatively less studies on GNN attributions, and their developments in drug research are still at the early stages. In this work, we adopted several advanced attribution techniques for different GNN frameworks and applied them to explain multiple drug molecule property prediction tasks, enabling the identification and visualization of vital chemical information in the networks. Additionally, we evaluated them quantitatively with attribution metrics such as accuracy, sparsity, fidelity and infidelity, stability and sensitivity; discussed their applicability and limitations; and provided an open-source benchmark platform for researchers. The results showed that all attribution techniques were effective, while those directly related to the predicted labels, such as integrated gradient, preferred to have better attribution performance. These attribution techniques we have implemented could be directly used for the vast majority of chemical GNN interpretation tasks.

An Automated Multi-Tab Website Fingerprinting Attack

In Website Fingerprinting (WF) attack, a local passive eavesdropper utilizes network flow information to identify which web pages a user is browsing. Previous researchers have demonstrated the feasibility and effectiveness of WF attacks under a strong Single Page Assumption: the network flow extracted by the adversary belongs to a single web page. In reality, the assumption may not hold because users tend to open multiple tabs simultaneously (or within a short period of time) so that their network traffic is mixed. In this article, we propose an automated multi-tab Website Fingerprinting attack that is able to accurately classify websites regardless of the number of simultaneously opened pages. Our design is powered by two innovative designs. First, we develop a split point classification method to dynamically identify the split point between the first page and its subsequent pages. As a result, the network traffic before the split point is solely generated for the first page. Then, we propose a new chunk-based WF classifier to infer the websites based on the initial chunk of clean traffic. For both classifiers, we apply automated feature selection to select a concise yet representative feature set. We implement a prototype of our design and perform extensive evaluations using SSH and Tor-based datasets to demonstrate the effectiveness of both our system components individually and the integrated system as a whole.

Asymmetric information flow in brain networks supporting expressive language in childhood.

Low-beta (13-23 Hz) event-related desynchrony (ERD), a neural signature of expressive language, lateralizes from bilateral to left hemisphere in development. In contrast, low-beta event-related synchrony (ERS), thought to reflect inhibition, lateralizes from bilateral to the right hemisphere across development. Using whole-brain directed connectivity analyses, we aimed to characterize hemispheric and regional contributions to expressive language, in childhood. We studied 80 children and adolescents, 4 to less than 19 years of age, performing covert auditory verb generation in magnetoencephalography. Outdegree, indegree, and betweenness centrality were used to differentiate regions acting as drivers, receivers, and bridging hubs, respectively. The number of suprathreshold connections significantly increased with age for delta band (p< .01). Delta outflow was mapped to left inferior frontal gyrus (IFG), while regions of right hemisphere, including right IFG, showed significant inflow. The right parietal cortex showed significant ERS, but without corresponding outdegree or indegree. Betweenness mapped to midline cortical and subcortical structures. Results suggest Broca's area develops a driving role in the language network, while Broca's homologue receives information without necessarily propagating it. Subcortical and midline hubs act as intrahemispheric relays. Findings suggest that Broca's homologue is inhibited during expressive language, in development.

Social networks and health information sharing in COVID-19 pandemic

Abstract Background Understanding health information flow in social networks is important for designing effective health communications strategies and to achieve health literacy. Limited information is known about variation in social networks and health information sharing in the COVID-19 pandemic by demographic factors. Young people are of particular interest given their heavy exposure to digital media sources, which include considerable health misinformation. Methods Hawai'i (n = 324) residents between 18-35 completed a Spring 2021 online survey including questions on health information flow in social networks: (1) how many they talked to and (2) listened to about health. Two Poisson regression models were fit with gender, education, and race/ethnicity predicting social network size. Results Respondents were 67.6% female. Respondents discussed their own health with 2-3 people (M = 2.18, SD = 2.95) and listened to roughly the same number. Respondents who talked with a greater number of individuals about their own health were significantly more likely to have larger networks for listening to others (r (317) = .614; p&amp;lt; .001). In the model for discussing their own health, as education increased so did social network size. For the model predicting discussing others’ health, gender was significant (p = 0.003); women listened to 30.6% more individuals than men. Most (73%) respondents had conducted a recent digital health search for either themselves or someone else, including for parents, grandparents, and children. Facebook (63%) and Instagram (58%) were the most popular online sources for COVID-19 health information. Conclusions Understanding social networks and digital health sources in young people are important for designing effective health communications to reach all communities, especially those experiencing health inequities, given the amount of health misinformation circulating and the need to build trust in public health communication. Key messages • Social networks provide access to critical health information including information obtained from digital sources. • Gender and education were important predictors of social network size in COVID-19 health communications.

Shining a Light in a Dark Corner: Does EDGAR Search Activity Reveal the Strategically Leaked Plans of Activist Investors?

AbstractWe provide evidence of a network of information flow between activists and other investors prior to 13D filings. We match EDGAR search activity to investor IP addresses, identifying specific investors who persistently download information on an individual activist’s campaign targets in the days prior to that activist’s 13D disclosures. This outside investor’s knowledge of pending activist campaign plans seems to benefit both parties: the informed investor, unnamed in the 13D, increases its holdings in the targeted stock prior to the price surge upon 13D disclosure, while the activist earns voting support that increases their likelihood of pursuing and winning a proxy fight.

Information flow estimation: A study of news on Twitter

News media has long been an ecosystem of creation, reproduction, and critique, where news outlets report on current events and add commentary to ongoing stories. Understanding the dynamics of news information creation and dispersion is important to accurately ascribe credit to influential work and understand how societal narratives develop. These dynamics can be modelled through a combination of information-theoretic natural language processing and networks; and can be parameterised using large quantities of textual data. However, it is challenging to see “the wood for the trees”, i.e., to detect small but important flows of information in a sea of noise. Here we develop new comparative techniques to estimate temporal information flow between pairs of text producers. Using both simulated and real text data we compare the reliability and sensitivity of methods for estimating textual information flow, showing that a metric that normalises by local neighbourhood structure provides a robust estimate of information flow in large networks. We apply this metric to a large corpus of news organisations on Twitter and demonstrate its usefulness in identifying influence within an information ecosystem, finding that average information contribution to the network is not correlated with the number of followers or the number of tweets. This suggests that small local organisations and right-wing organisations which have lower average follower counts still contribute significant information to the ecosystem. Further, the methods are applied to smaller full-text datasets of specific news events across news sites and Russian troll accounts on Twitter. The information flow estimation reveals and quantifies features of how these events develop and the role of groups of trolls in setting disinformation narratives. In summary, this work provides a new methodology for examining the information transmitted between content producers in any connected system of natural language, a toolkit with applications to the many networked discourses of our online world.