Network Topology Research Articles

Human-connected wild lands: How network analysis can inform invasive species management

Wild landscapes are critical strongholds for biodiversity, yet even in the remotest parts of the globe, increasing human use and development are leading to an influx of biodiversity threats including invasive species. Natural resource management agencies, and those that rely on public lands, need a better understanding of the long-distance dispersal pathways in which invasive species are introduced to remote locations. Pathway information is essential for targeting prevention and early detection across vast landscapes, but it is often challenged by information gaps and high surveillance costs. Data-driven approaches centered around a participating public can help resource managers and biosecurity professionals to better prioritize prevention and early detection activities to minimize incipient and secondary invasions. This study employed surveys with resource users to integrate and analyze multiple human-mediated dispersal networks for aquatic invasive species (AIS) across Alaska's part of the North American Boreal Forest. Specifically, it applied network analysis to further inform management priorities that so far were only based on a single pathway and different metrics. Results underline the vulnerability of remote wild freshwater systems to the introduction of AIS and provide a waterbody-specific tool for prioritizing monitoring and inform pathway-specific interventions that were unavailable through past research. The study compares the prioritization of waterbodies under newly derived network topology metrics accounting for multiple generic pathways with the existing single- species and single-pathway prediction model. Advantages of a more comprehensive multi-pathway network topology are discussed in the context of various invasion stages, multiple invasive taxa, and resource constraint conservation management systems.

Atomistic insight into structure and properties of oxide films formed upon oxidation of Al–Mg alloys – reactive molecular dynamics study

Oxidation phenomena on metal surfaces can be advanced using atomistic modeling for the study of the structure and properties of the growing oxide films. Molecular dynamics investigations were performed to study thermal oxidation of single-crystal and polycrystalline Al–Mg alloys with low Mg content of up to 2.5 at. %. Structure, topological atom network and surface topography of the oxide films grown on Al–Mg surfaces at 300, 475 and 663 K are determined. Mg-content and temperature dependent oxide films developed on Al-Mg alloy substrates with two-phase oxide film formation as also confirmed experimentally. It is related with a gradual long-range order formation above 475 K for (Al,Mg)-oxide films, for which solid amorphous phase predominates over crystalline phase. The 1.12-nm-thick oxide films grown at 300 K are fully amorphous and build up from (Al, Mg)-oxide which the Al2O3 dominates. Increase of Mg coefficient confirms faster Mg diffusion to oxide/alloy interface at high oxidation temperature and the formation of MgO-like network. The presence of planar crystal defect (grain boundary, GB) into Al-Mg alloy substrate governs the kinetics of oxide film growth and its structure evolution. GB also compensate internal stresses during oxide film growth. Obtained results are in good agreement with experimental observations.

Functional activity, functional connectivity and complex network biomarkers of progressive hyposmia Parkinson's disease with no cognitive impairment: evidences from resting-state fMRI study.

Olfactory dysfunction stands as one of the most prevalent non-motor symptoms in the initial stage of Parkinson's disease (PD). Nevertheless, the intricate mechanisms underlying olfactory deficits in Parkinson's disease still remain elusive. This study collected rs-fMRI data from 30 PD patients [15 with severe hyposmia (PD-SH) and 15 with no/mild hyposmia (PD-N/MH)] and 15 healthy controls (HC). To investigate functional segregation, the amplitude of low-frequency fluctuation (ALFF) and regional homogeneity (ReHo) were utilized. Functional connectivity (FC) analysis was performed to explore the functional integration across diverse brain regions. Additionally, the graph theory-based network analysis was employed to assess functional networks in PD patients. Furthermore, Pearson correlation analysis was conducted to delve deeper into the relationship between the severity of olfactory dysfunction and various functional metrics. We discovered pronounced variations in ALFF, ReHo, FC, and topological brain network attributes across the three groups, with several of these disparities exhibiting a correlation with olfactory scores. Using fMRI, our study analyzed brain function in PD-SH, PD-N/MH, and HC groups, revealing impaired segregation and integration in PD-SH and PD-N/MH. We hypothesize that changes in temporal, frontal, occipital, and cerebellar activities, along with aberrant cerebellum-insula connectivity and node degree and betweenness disparities, may be linked to olfactory dysfunction in PD patients.

GNN4DM: a graph neural network-based method to identify overlapping functional disease modules.

Identifying disease modules within molecular interaction networks is an essential exploratory step in computational biology, offering insights into disease mechanisms and potential therapeutic targets. Traditional methods often struggle with the inherent complexity and overlapping nature of biological networks, and they are limited in effectively leveraging the vast amount of available genomic data and biological knowledge. This limitation underscores the need for more effective, automated approaches to integrate these rich data sources. In this work, we propose GNN4DM, a novel graph neural network-based structured model that automates the discovery of overlapping functional disease modules. GNN4DM effectively integrates network topology with genomic data to learn the representations of the genes corresponding to functional modules and align these with known biological pathways for enhanced interpretability. Following the DREAM benchmark evaluation setting and extending with three independent data sources (GWAS Atlas, FinnGen, and DisGeNET), we show that GNN4DM performs better than several state-of-the-art methods in detecting biologically meaningful modules. Moreover, we demonstrate the method's applicability by discovering two novel multimorbidity modules significantly enriched across a diverse range of seemingly unrelated diseases. Source code, all training data, and all identified disease modules are freely available for download at https://github.com/gezsi/gnn4dm. GNN4DM was implemented in Python.

Molecular mechanism of Dang-Shen-Yu-Xing decoction against Mycoplasma bovis pneumonia based on network pharmacology, molecular docking, molecular dynamics simulations and experimental verification.

Mycoplasma bovis pneumonia is a highly contagious respiratory infection caused by Mycoplasma bovis. It is particularly prevalent in calves, posing a significant threat to animal health and leading to substantial economic losses. Dang-Shen-Yu-Xing decoction is often used to treat this condition in veterinary clinics. It exhibits robust anti-inflammatory effects and can alleviate pulmonary fibrosis. However, its mechanism of action remains unclear. Therefore, this study aimed to preliminarily explore the molecular mechanism of Dang-Shen-Yu-Xing decoction for treating mycoplasma pneumonia in calves through a combination of network pharmacology, molecular docking, molecular dynamics simulation methods, and experimental validation. The active components and related targets of Dang-Shen-Yu-Xing decoction were extracted from several public databases. Additionally, complex interactions between drugs and targets were explored through network topology, Gene Ontology, and Kyoto Encyclopedia of Genes and Genomes enrichment analyses. Subsequently, the binding affinity of drug to disease-related targets was verified through molecular docking and molecular dynamics simulation. Finally, the pharmacodynamics were verified via animal experiments. The primary network topology analysis revealed two core targets and 10 key active components of Dang-Shen-Yu-Xing decoction against Mycoplasma bovis pneumonia. Kyoto Encyclopedia of Genes and Genomes enrichment analysis showed that the mechanism of Dang-Shen-Yu-Xing decoction for treating mycoplasma bovis pneumonia involved multiple signaling pathways, with the main pathways including PI3K-Akt and IL17 signaling pathways. Moreover, molecular docking predicted the binding affinity and conformation of the core targets of Dang-Shen-Yu-Xing decoction, IL6, and IL10, with the associated main active ingredients. The results showed a strong binding of the active ingredients to the hub target. Further, molecular docking dynamics simulation revealed three key active components of IL10 induced by Dang-Shen-Yu-Xing decoction against Mycoplasma bovis pneumonia. Finally, animal experiments confirmed Dang-Shen-Yu-Xing decoction pharmacodynamics, suggesting that it holds potential as an alternative therapy for treating mycoplasma bovis pneumonia.

Baicalin Ameliorates Depression-like Behaviors via Inhibiting Neuroinflammation and Apoptosis in Mice

Depression is a common neuropsychiatric disease which brings an increasing burden to all countries globally. Baicalin, a flavonoid extracted from the dried roots of Scutellaria, has been reported to exert anti-inflammatory, antioxidant, and neuroprotective effects in the treatment of depression. However, the potential biological mechanisms underlying its antidepressant effect are still unclear. In the present study, we conducted extensive research on the potential mechanisms of baicalin’s antidepressant effect using the methods of network pharmacology, including overlapped terms-based analysis, protein–protein interaction (PPI) network topology analysis, and enrichment analysis. Moreover, these results were further verified through molecular docking, weighted gene co-expression network analysis (WGCNA), differential gene expression analysis, and subsequent animal experiments. We identified forty-one genes as the targets of baicalin in the treatment of depression, among which AKT1, IL6, TP53, IL1B, and CASP3 have higher centrality in the more core position. Meanwhile, the roles of peripheral genes derived from direct potential targets were also observed. Our study suggested that biological processes, such as inflammatory reaction, apoptosis, and oxidative stress, may be involved in the therapeutic process of baicalin on depression. These mechanisms were validated at the level of structure, gene, protein, and signaling pathway in the present study. Taken together, these findings propose a new perspective on the potential mechanisms underlying baicalin’s antidepressant effect, and also provide a new basis and clarified perspective for its clinical application.

Characterization of the α-relaxation and gel-like dynamics in mixed network forming (GeO2)x(NaPO3)100-x glass melts by photon correlation spectroscopy

Dynamic light scattering investigations of the α-relaxation in network-forming (GeO2)x(NaPO3)100-x glass melts is reported for x = 0 to 40 mol% GeO2. Addition of GeO2 results in an increased density of bridging oxygen bonds. This monotonically increases the glass transition temperature but generates complex changes to the glass forming fragility which initially decreases with increasing GeO2 but increases again at higher GeO2 contents. A resolution to this odd fragility behavior is obtained by applying a coarse-graining procedure used previously in sodium germanate melts to arrive at a topological network connectivity for which fragility exhibits a universal dependence. Also observed in the light scattering measurements is a secondary relaxation, several orders of magnitude slower than that of the α-relaxation, which we tentatively attribute to gel-like anomalous diffusion of Ge crosslinks formed between phosphate chains.

An innovative NSGA-II-based Byzantine Fault Tolerant solution for software defined network environments

Byzantine fault tolerance (BFT) of the control plane in Software Defined Networking (SDN) is achieved by mapping each switch to 3f+1 number of controllers, where f represents the number of faulty controllers that can be tolerated at a time. A BFT approach protects the data plane from any potential malicious activity at the control plane by detecting the inconsistency among the response messages from multiple controllers. To compute the optimal mapping of switches to the controller, the existing literature does not consider some important parameters. This paper proposes a novel approach, named NBFT-SDN, that extends an artificial intelligence algorithm (i.e. NSGA-II) to solve a new formulated multi-objective optimization problem associated with this mapping. NBFT-SDN considers the very important parameters link reliability and link load along with switch-to-controller minimum delay, switch-to-controllers maximum reliability, controller-to-controller minimum delay, minimum link load, minimum hop count, and controller load balancing when mapping the switches to the controllers in optimum manner. The performance of our proposed approach is evaluated in comparison to a state-of-art approach using real network traces with network topologies of diverse sizes. Our proposed approach NBFT-SDN show improved network performance in terms of reliability, delay, hop count, load balancing and link load.

Unlocking Auxetic Behavior in Recyclable Thermosetting Foams Enabled by Dynamic Disulfide Cross‐linking Strategy

AbstractAuxetic foams with a negative Poisson's ratio (NPR) have attracted considerable attention in material engineering due to their outstanding performance in seismic and energy absorption. Nevertheless, thermoplastic auxetic foams are compromised by weak non‐covalent crosslinking that diminishes the mechanical strength and durability of foams. Conversely, thermosetting foams with chemical crosslinking, although mechanically robust, face challenges in elaborating auxetic structure and in achieving recyclability. Herein, an alternative approach is proposed to tackle this dilemma by incorporating dynamic disulfide bonds into the polymer network for preparing a thermosetting polyurethane foam with covalent adaptable network. By leveraging the unidirectional multi‐effect compression technique, the topological network reorganization of foam is induced, transforming the initial circular open‐cell structure into a re‐entrant cell structure. This structural transformation endows the foam with stable NPR capability, achieving a minimum Poisson's ratio value of −0.4 within 30% compressive strain. Benefiting from its reinforced network structure, the foam also demonstrates high compressive strength (6.47 MPa) and tensile strength (1.67 MPa). Furthermore, it is recyclable and can be recompressed into thermosetting films. This work offers a straightforward approach to making auxetic thermosetting foams with good mechanical and recyclable properties, which is interesting for the development of high‐performance auxetic materials.

Analyzing the importance of network topology in AADT estimation: insights from travel demand models using graph neural networks

Analyzing the importance of network topology in AADT estimation: insights from travel demand models using graph neural networks

Impact of network topology changes on information source localization

Well-established methods of locating the source of information in a complex network are usually derived with the assumption of complete and exact knowledge of network topology. We study the performance of three such algorithms (Limited Pinto–Thiran–Vetterli Algorithm — LPTVA, Gradient Maximum Likelihood Algorithm — GMLA and Pearson Correlation Algorithm — PCA) in scenarios that do not fulfill this assumption by modifying the network before localization. This is done by adding superfluous new links, hiding existing ones, or reattaching links following the network’s structural Hamiltonian. Our results show that GMLA is highly resilient to adding superfluous edges, as its precision falls by more than statistical uncertainty only when the number of links is approximately doubled. On the other hand, if the edge set is underestimated or reattachment has taken place, the performance of GMLA drops significantly. In such a scenario, PCA is preferable, retaining most of its performance when other simulation parameters favor successful localization (high density of observers, highly deterministic propagation). It is also generally more accurate than LPTVA and orders of magnitude faster. The differences between localization algorithms can be intuitively explained, although further theoretical research is needed.

Theoretical insights into the diverse stabilizing effects of fenchyl alcohol/thymol eutectic solvent on carotenoids

Carotenoids, a class of vital natural pigments, are known for their antioxidant properties and are integral to various industrial applications. However, their instability during processing and storage has limited their broader utilization. This study investigated the stabilizing effects of a fenchyl alcohol/thymol-based deep eutectic solvent (DES) on a spectrum of carotenoids, including lutein, zeaxanthin, astaxanthin, and β-carotene. Through a combination of ab initio molecular dynamics (AIMD) simulations and quantum chemical calculations, we elucidated the molecular interactions and stabilization mechanisms within the DES environment (methanol as the benchmark). Our findings reveal that the DES could exert varying degrees of influence on the stability of different types of carotenoids. The temperature-dependent behavior of carotenoids in the DES suggested a complex interplay between hydrogen bonding, dispersion forces, and the inherent structural properties of carotenoids. AIMD simulations provided a detailed picture of the structural organization and hydrogen bond network topology within the DES, highlighting the role of temperature on the stability and orientation of solvent molecules around carotenoids. Energy decomposition analysis (EDA), Hirshfeld surface analysis, and visual study of interactions further dissected the non-covalent interactions, emphasizing the significance of electrostatic and dispersion forces-driven non-covalent interactions (van der Waals dispersion interactions) in stabilizing carotenoids within the DES matrix. In comparison to methanol, the interactions between components of the eutectic system and the target molecules demonstrate enhanced dynamism, increased complementarity, and a greater variety of interaction patterns. This research presented a novel perspective on the use of DESs for the stabilization of carotenoids, offering potential solutions to enhance their industrial applicability and performance. The insights gained from this study are crucial for the development of sustainable and eco-friendly solvent systems, aligning with the growing demand for green chemistry in the pharmaceutical, cosmetic, and food industries.

Abnormal Structural–Functional Coupling and MRI Alterations of Brain Network Topology in Progressive Supranuclear Palsy

BackgroundProgressive supranuclear palsy (PSP) can cause structural and functional brain reconstruction. There is a lack of knowledge about the consistency between structural–functional (S–F) connection networks in PSP, despite growing evidence of anomalies in various single brain network parameters.PurposeTo study the changes in the structural and functional networks of PSP, network's topological properties including degree, and the consistency of S–F coupling. The relationship with clinical scales was examined including the assessment of PSP severity, and so on.Study TypeRetrospective.SubjectsA total of 51 PSP patients (70.04 ± 7.46, 25 females) and 101 healthy controls (64.58 ± 8.84, 58 females).Field Strength/Sequence3‐T, resting‐state functional MRI, diffusion tensor imaging, and T1‐weighted images.AssessmentA graph‐theoretic approach was used to evaluate structural and functional network topology metrics. We used the S–F coupling changes to explore the consistency of structural and functional networks.Statistical TestsIndependent samples t tests were employed for continuous variables, χ2 tests were used for categorical variables. For network analysis, two‐sample t tests was used and implied an false discovery rate (FDR) correction. Pearson correlation analysis was used to explore the correlations. A P‐value <0.05 was considered statistically significant.ResultsPSP showed variations within and between modules. Specifically, PSP had decreased network properties changes (t = −2.0136; t = 2.5409; t = −2.5338; t = −2.4296; t = −2.5338; t = 2.8079). PSP showed a lower coupling in the thalamus and left putamen and a higher coupling in the visual, somatomotor, dorsal attention, and ventral attention network. S–F coupling was related to the number of network connections (r = 0.32, r = 0.22) and information transmission efficiency (r = 0.55, r = 0.28). S–F coupling was related to basic academic ability (r = 0.39) and disinhibition (r = 0.49).Data ConclusionPSP may show abnormal S–F coupling and intramodular and intermodular connectome in the structural and functional networks.Level of Evidence3Technical EfficacyStage 3

Decentralized Mechanism for Edge Node Allocation in Access Network: An Experimental Evaluation

With the rapid advancement of the Internet of Things and the emergence of 6G networks in smart city environments, a growth in the generation of data, commonly known as big data, is expected to consequently lead to higher latency. To mitigate this latency, mobile edge computing has been proposed to alleviate a portion of the workload from mobile devices by offloading it to nearby edge servers equipped with appropriate computational resources. However, existing solutions often exhibit poor performance when confronted with complex network topologies. Thus, this paper introduces a decentralized mechanism aimed at determining the locations of network edge nodes in such complex network topologies, characterized by lengthy execution times. Our proposal provides performance improvements and offers scalability and flexibility as networks become more complex. Experimental evaluations are conducted using the Shanghai Telecom dataset to validate our proposed approach.

Distributed associations among white matter hyperintensities and structural brain networks with fluid cognition in healthy aging

White matter hyperintensities (WMHs) are associated with age-related cognitive impairment and increased risk of Alzheimer’s disease. However, the manner by which WMHs contribute to cognitive impairment is unclear. Using a combination of predictive modeling and network neuroscience, we investigated the relationship between structural white matter connectivity and age, fluid cognition, and WMHs in 68 healthy adults (18–78 years). Consistent with previous work, WMHs were increased in older adults and exhibited a strong negative association with fluid cognition. Extending previous work, using predictive modeling, we demonstrated that age, WMHs, and fluid cognition were jointly associated with widespread alterations in structural connectivity. Subcortical-cortical connections between the thalamus/basal ganglia and frontal and parietal regions of the default mode and frontoparietal networks were most prominent. At the network level, both age and WMHs were negatively associated with network density and communicability, and positively associated with modularity. Spatially, WMHs were most prominent in arterial zones served by the middle cerebral artery and associated lenticulostriate branches that supply subcortical regions. Finally, WMHs overlapped with all major white matter tracts, most prominently in tracts that facilitate subcortical-cortical communication and are implicated in fluid cognition, including the anterior thalamic-radiations and forceps minor. Finally, results of mediation analyses suggest that whole-brain WMH load influences age-related decline in fluid cognition. Thus, across multiple levels of analysis, we showed that WMHs were increased in older adults and associated with altered structural white matter connectivity and network topology involving subcortical-cortical pathways critical for fluid cognition.

Adaptive neural network based quadrotor UAV formation control under external disturbances

The formation control of a team comprised of multiple quadrotor Unmanned Aerial Vehicles (UAVs) may severely be affected by the unknown external disturbances. The external disturbances are caused by wind forces to create aero-dynamical disturbances. This article addresses the robust formation control problem of multiple UAVs system despite the effect of external disturbances that allow sustaining a stable network connection among the UAVs and maintaining different formations assigned to them. First, a Radial Basis Function Neural Network (RBFNN) based model is developed to reciprocate the external disturbances along the positional and the attitude subsystems. Then incorporating the estimated disturbance values a distributed adaptive formation controller is devised using the Lyapunov theory. It consists of a positional and an attitude controller associated with the translational and the rotational movements of the UAVs. The stability is validated by satisfying the criteria of the Lyapunov stability function. The UAVs are connected through variable adjacency matrix based directed network topology and the network connectivity is established through the properties of the Laplacian Matrix. The robustness of the designed controller is justified via rigorous simulation studies for different sets of desired formations such as triangular, squared, tetrahedron, octahedron and cube shaped. The reference trajectories are considered as spiral, straight line and circular shaped. The time varying external disturbances are considered of sinusoidal waveform of different magnitudes. The simulation results signifies that the proposed RBFNN based formation controller reciprocate different sinusoidal waveforms to achieve the desired formations successfully. Extensive comparative studies demonstrate the efficacy of the proposed adaptive formation controller over the existing controllers presented in the literature for different shapes of trajectories and desired formations.

Multifunctional Nano‐Conductive Hydrogels With High Mechanical Strength, Toughness and Fatigue Resistance as Self‐Powered Wearable Sensors and Deep Learning‐Assisted Recognition System

AbstractHigh mechanical strength, toughness, and fatigue resistance are essential to improve the reliability of conductive hydrogels for self‐powered sensing. However, achieving mutually exclusive properties simultaneously remains challenging. Hence, a novel directed interlocking strategy based on topological network structure and mechanical training is proposed to construct tough hydrogels by optimizing the network structure and modulating the orientation of molecular chains. Combining Zn2+ crosslinked cellulose nanofibers (CNFs) and a polyacrylamide‐poly(vinyl alcohol) double‐network, the unique interlocked‐network structure exhibits an enhanced toughening effect due to hydrogen bonding and metal‐ligand interactions. The aligned nanocrystalline domains achieved by training further contribute to an increase in the toughness and fatigue thresholds. This innovative approach synergistically enhances the mechanical properties of the nano‐conductive hydrogel, achieving a maximum tensile strength of 4.98 MPa and a toughness of 48 MJ m−3. Notably, the CNFs template with anchored polyaniline, when oriented through mechanical training, forms a unique directional conductive pathway, which significantly enhances the power output performance. Besides, a motion recognition system based on a self‐powered sensing device is designed with the assistance of deep learning techniques to accurately identify human motion behaviors. This work showcases a potentially transformative flexible electronic material for self‐powered sensing systems and intelligent recognition systems.

Hybrid optimization-based secured routing in mobile ad-hoc network

Mobile Ad hoc Networks (MANETs) pose significant routing challenges due to their decentralized and dynamic nature. Node mobility results in frequent changes in network topology, leading to unsFIG connectivity and link quality. Traditional RPs designed for static networks are inadequate for MANETs. To deal with these issues, a secure routing approach is proposed using the Red Panda-Lyrebrid Optimization (RePLO) algorithm, which combines the advantages of the Red Panda Optimization (RPO) and Lyrebird Optimization Algorithm (LOA) algorithms. The proposed approach consists of five steps: (i) configuring the system model, (ii) developing the energy model, (iii) creating the mobility model, (iv) selecting cluster heads using the RePLO algorithm, and (v) routing using the RePLO algorithm. The RePLO algorithm optimizes cluster head selection and routing while considering specific constraints such as delay, distance, energy, & security for Cluster Head (CH) selection, and link quality and enhanced trust for routing optimization. The effectiveness of the proposed approach is evaluated using various metrics to demonstrate its efficiency in MANET routing. By integrating multiple optimization techniques and considering critical constraints, the RePLO algorithm offers a systematic and secure solution for MANET routing. The evaluation results confirm the efficacy of the proposed approach in improving network performance, reliability, and security. Overall, the RePLO algorithm presents a promising approach to tackle the routing issues inherent in MANETs, paving the way for more robust and efficient communication in mobile ad hoc networks.

Assessment of Solubility Behavior of a Copolymer in Supercritical CO2 and Organic Solvents: Neural Network Prediction and Statistical Analysis.

In the industrial sector, understanding the behavior of block copolymers in supercritical solvents is crucial. While qualitative agreement with polymer solubility curves has been evaluated using complex theoretical models in many cases, quantitative predictions remain challenging. This study aimed to create a rapid and accurate artificial neural network (ANN) model to predict the lower critical solubility and upper critical solubility space of an atypical block copolymer, poly(styrene-co-octafluoropentyl methacrylate) (PSOM), in different supercritical solvent systems over a wide range of temperatures (51.75-182.05 °C) and high pressure (3.28-200.86 MPa). The experimental data set used in this study included one copolymer, five supercritical solvents, one cosolvent, and one initiator. It consisted of seven unique copolymer-solvent combinations (252 cloud point pressures) used to predict the model quantitatively and qualitatively. To predict the PSOM-solvent interactions, the study considered two different input systems: a six-variable system, a five-variable system, and one target output. Initially, we used a three-layer feed-forward neural network to select the best learning algorithm (Levenberg-Marquardt) from 14 different algorithms, considering one sample PSOM-solvent system. Then, the network topology was optimized by varying hidden neuron numbers from 2 to 80 for all seven PSOM-solvent combination systems. The predicted cloud point pressures were in excellent agreement with the experimental cloud point pressures, confirming the model's accuracy. It is clear from the results of a minimum mean square error (≤1.90 × 10-27) and maximum linear regression R 2 (≥0.99) during training, validation, and testing of all the data sets. Further, the ANN model accuracy was tested by statistical analysis, confirming the model's ability to accurately capture the miscibility regions of polymers, enabling efficient processing of various polymer materials. This data-driven approach facilitates the prediction of coexistence curves for other polymers and complex macromolecular systems.

InfoFlowNet: A multi-head attention-based self-supervised learning model with surrogate approach for uncovering brain effective connectivity

Deciphering brain network topology can enhance the depth of neuroscientific knowledge and facilitate the development of neural engineering methods. Effective connectivity, which gauges the directional influences among brain regions, is pivotal for these studies. This research introduces InfoFlowNet, a novel self-supervised learning model designed to infer causal relationships in electroencephalogram (EEG) data, aiming at capturing the complex information flow among brain regions. Specifically, InfoFlowNet employs convolution operations and multi-head self-attention mechanisms with a masking feature, which ignores self-connections to focus on inter-regional dynamics. Additionally, a new causal magnitude estimation is introduced by assessing the impact of shuffled surrogate data on signal prediction to quantify causal influence. Experiments using synthetic data and real EEG datasets demonstrate that InfoFlowNet effectively uncovers time-varying causal relationships. Compared with the Granger causality model (GCM) and the temporal causal discovery framework (TCDF), InfoFlowNet shows superior sensitivity in detecting significant causal edges. For instance, in a psychomotor vigilance task, InfoFlowNet identified 8 out of 16 significant causal edges, significantly outperforming GCM and TCDF. Furthermore, InfoFlowNet successfully passes the whiteness test on the model's residuals, and advanced analyses illustrate the advantages of utilizing multi-head attention and masking mechanisms to capture effective connectivity. In sum, InfoFlowNet represents a significant advancement in the analysis of effective connectivity, offering a deeper understanding of brain network dynamics. The model's ability to discern complex causal interactions supports its potential application in broader neuroscientific research and applications.