BISoN: A Bayesian Framework for Inference of Social Networks

Animal social networks are often constructed from point estimates of edge weights. In many contexts, edge weights are inferred from observational data, and the uncertainty around estimates can be affected by various factors. Though this has been acknowledged in previous work, methods that explicitly quantify uncertainty in edge weights have not yet been widely adopted, and remain undeveloped for many common types of data. Furthermore, existing methods are unable to cope with some of the complexities often found in observational data, and do not propagate uncertainty in edge weights to subsequent statistical analyses.We introduce a unified Bayesian framework for modelling social networks based on observational data. This framework, which we call BISoN, can accommodate many common types of observational social data, can capture confounds and model effects at the level of observations, and is fully compatible with popular methods used in social network analysis.We show how the framework can be applied to common types of data and how various types of downstream statistical analyses can be performed, including non-random association tests and regressions on network properties.Our framework opens up the opportunity to test new types of hypotheses, make full use of observational datasets, and increase the reliability of scientific inferences. We have made both an R package and example R scripts available to enable adoption of the framework.

Sampling bias and the robustness of ecological metrics for plant-damage-type association networks

Plants and their insect herbivores have been a dominant component of the terrestrial ecological landscape for the past 410 million years and feature intricate evolutionary patterns and co-dependencies. A complex systems perspective allows for both detailed resolution of these evolutionary relationships as well as comparison and synthesis across systems. Using proxy data of insect herbivore damage (denoted by the damage type or DT) preserved on fossil leaves, functional bipartite network representations provide insights into how plant-insect associations depend on geological time, paleogeographical space, and environmental variables such as temperature and precipitation. However, the metrics measured from such networks are prone to sampling bias. Such sensitivity is of special concern for plant-DT association networks in paleontological settings where sampling effort is often severely limited. Here, we explore the sensitivity of functional bipartite network metrics to sampling intensity and identify sampling thresholds above which metrics appear robust to sampling effort. Across a broad range of sampling efforts, we find network metrics to be less affected by sampling bias and/or sample size than richness metrics, which are routinely used in studies of fossil plant-DT interactions. These results provide reassurance that cross-comparisons of plant-DT networks offer insights into network structure and function and support their widespread use in paleoecology. Moreover, these findings suggest novel opportunities for using plant-DT networks in neontological terrestrial ecology to understand functional aspects of insect herbivory across geological time, environmental perturbations, and geographic space.

Connectivity in eQTL networks dictates reproducibility and genomic properties

Expression quantitative trait locus (eQTL) analysis associates SNPs with gene expression; these relationships can be represented as a bipartite network with association strength as "edge weights" between SNPs and genes. However, most eQTL networks use binary edge weights based on thresholded FDR estimates: definitions that influence reproducibility and downstream analyses. We constructed twenty-nine tissue-specific eQTL networks using GTEx data and evaluated a comprehensive set of network specifications based on false discovery rates, test statistics, and p values, focusing on the degree centrality-a metric of an SNP or gene node's potential network influence. We found a thresholded Benjamini-Hochberg q value weighted by the Z-statistic balances metric reproducibility and computational efficiency. Our estimated gene degrees positively correlate with gene degrees in gene regulatory networks, demonstrating that these networks are complementary in understanding regulation. Gene degrees also correlate with genetic diversity, and heritability analyses show that highly connected nodes are enriched for tissue-relevant traits.

Microstructural Characteristics of the Weighted and Directed International Crop Trade Networks

With increasing global demand for food, international food trade is playing a critical role in balancing the food supply and demand across different regions. Here, using trade datasets of four crops that provide more than 50% of the calories consumed globally, we constructed four international crop trade networks (iCTNs). We observed the increasing globalization in the international crop trade and different trade patterns in different iCTNs. The distributions of node degrees deviate from power laws, and the distributions of link weights follow power laws. We also found that the in-degree is positively correlated with the out-degree, but negatively correlated with the clustering coefficient. This indicates that the numbers of trade partners affect the tendency of economies to form clusters. In addition, each iCTN exhibits a unique topology which is different from the whole food network studied by many researchers. Our analysis on the microstructural characteristics of different iCTNs provides highly valuable insights into distinctive features of specific crop trades and has potential implications for model construction and food security.

Mapping Structural Connectivity Using Diffusion MRI: Challenges and Opportunities

Diffusion MRI-based tractography is the most commonly-used technique when inferring the structural brain connectome, i.e., the comprehensive map of the connections in the brain. The utility of graph theory-a powerful mathematical approach for modeling complex network systems-for analyzing tractography-based connectomes brings important opportunities to interrogate connectome data, providing novel insights into the connectivity patterns and topological characteristics of brain structural networks. When applying this framework, however, there are challenges, particularly regarding methodological and biological plausibility. This article describes the challenges surrounding quantitative tractography and potential solutions. In addition, challenges related to the calculation of global network metrics based on graph theory are discussed.Evidence Level: 5Technical Efficacy: Stage 1.

Connectivity within regions characterizes epilepsy duration and treatment outcome

Finding clear connectome biomarkers for temporal lobe epilepsy (TLE) patients, in particular at early disease stages, remains a challenge. Currently, the whole-brain structural connectomes are analyzed based on coarse parcellations (up to 1,000 nodes). However, such global parcellation-based connectomes may be unsuitable for detecting more localized changes in patients. Here, we use a high-resolution network (~50,000-nodes overall) to identify changes at the local level (within brain regions) and test its relation with duration and surgical outcome. Patients with TLE (n = 33) and age-, sex-matched healthy subjects (n = 36) underwent high-resolution (~50,000 nodes) structural network construction based on deterministic tracking of diffusion tensor imaging. Nodes were allocated to 68 cortical regions according to the Desikan-Killany atlas. The connectivity within regions was then used to predict surgical outcome. MRI processing, network reconstruction, and visualization of network changes were integrated into the NICARA (https://nicara.eu). Lower clustering coefficient and higher edge density were found for local connectivity within regions in patients, but were absent for the global network between regions (68 cortical regions). Local connectivity changes, in terms of the number of changed regions and the magnitude of changes, increased with disease duration. Local connectivity yielded a better surgical outcome prediction (Mean value: 95.39% accuracy, 92.76% sensitivity, and 100% specificity) than global connectivity. Connectivity within regions, compared to structural connectivity between brain regions, can be a more efficient biomarker for epilepsy assessment and surgery outcome prediction of medically intractable TLE.

Environmental Display Can Buffer the Effect of Pesticides on Solitary Bees

Environmental quality (e.g., diversity of resource availability, nesting sites, environmental display) plays an important role in an animal’s life. While homogeneous environments can restrict organisms from developing activities such as food seeking (behavioral impairment), more complex environments allow animals to perform activities with learning and behavioral perfecting outcomes. Pesticides are known to affect the learning and foraging behaviors of bees; however, little is known about the counterbalance displayed by the environment. Herein, we conducted two experiments that simulated distinct environmental displays, in which the effects of a fungicide (Indar^TM 5EW-febunconazole) on solitary bee foraging activities were tested. We found that the fungicide only impaired the activities of bees in one of the studied environments. The difference in visitation rates and flower exploitation of bees between the two different environmental displays led to changes in metrics of bee–flower networks across environments. Linkage density, a metric associated with pollination efficiency that is known to be impacted by different environments, differed across environments. Our results showed that ecological interaction network metrics can differ regarding the different environmental displays. This study indicates that environmental complexity helps balance the negative effects of pesticides on solitary bees and highlights the potential use of solitary bees as model organisms for experimental simulations of environmental change.

Finding missing links in interaction networks

Documenting which species interact within ecological communities is challenging and labor intensive. As a result, many interactions remain unrecorded, potentially distorting our understanding of network structure and dynamics. We test the utility of four structural models and a new coverage-deficit model for predicting missing links in both simulated and empirical bipartite networks. We find they can perform well, although the predictive power of structural models varies with the underlying network structure. The accuracy of predictions can be improved by ensembling multiple models. Augmenting observed networks with most-likely missing links improves estimates of qualitative network metrics. Tools to identify likely missing links can be simple to implement, allowing the prioritization of research effort and more robust assessment of network properties.

Using social-network research to improve outcomes in natural resource management

The conservation and management of natural resources operates in social-ecological systems in which resource users are embedded in social and environmental contexts that influence their management decisions. Characterizing social networks of resource users can be used to inform understanding of social influences on decision making, and social network analysis (SNA) has emerged as a useful technique to explore these relationships. We synthesized how SNA has been used in 85 studies of natural resource management. We considered how social networks and social processes (e.g., interactions between individuals) influence each other and in turn influence social outcomes (e.g., decisions or actions) that affect environmental outcomes (e.g., improved condition). Descriptive methods were used in 58% of the studies to characterize social processes, and 42% of the studies compared multiple networks or multiple points in time to assess social or environmental outcomes. In 4 studies, authors assessed network interventions intended to affect social processes or environmental outcomes. The heterogeneity in case studies, methods, and analyses preclude general lessons. Thus, to structure and further learning about the role of social networks in achieving environmental outcomes, we created a typology that deconstructs social processes, social outcomes, and environmental outcomes into themes and options of social and ecological measures within each. We suggest shifts in research foci toward intervention studies to aid in understanding causality and inform the design of conservation initiatives. There is a need to develop clearer justification and guidance around the proliferation of network measures. The use of SNA in natural resource management is expanding rapidly; thus, now is the time for the conservation community to build a more rigorous evidence base to demonstrate the extent to which social networks can play a role in achieving desired social and environmental outcomes.

Knowledgeable Lemurs Become More Central in Social Networks

Strong relationships exist between social connections and information transmission [1-9], where individuals' network position plays a key role in whether or not they acquire novel information [2, 3, 5, 6]. The relationships between social connections and information acquisition may be bidirectional if learning novel information, in addition to being influenced by it, influences network position. Individuals who acquire information quickly and use it frequently may receive more affiliative behaviors [10, 11] and may thus have a central network position. However, the potential influence of learning on network centrality has not been theoretically or empirically addressed. To bridge this epistemic gap, we investigated whether ring-tailed lemurs' (Lemur catta) centrality in affiliation networks changed after they learned how to solve a novel foraging task. Lemurs who had frequently initiated interactions and approached conspecifics before the learning experiment were more likely to observe and learn the task solution. Comparing social networks before and after the learning experiment revealed that the frequently observed lemurs received more affiliative behaviors than they did before-they became more central after the experiment. This change persisted even after the task was removed and was not caused by the observed lemurs initiating more affiliative behaviors. Consequently, quantifying received and initiated interactions separately provides unique insights into the relationships between learning and centrality. While the factors that influence network position are not fully understood, our results suggest that individual differences in learning and becoming successful can play a major role in social centrality, especially when learning from others is advantageous.

Node self-connections and metapopulation persistence: reply to Saura (2018)

Saura () claims that studies using the Probability of Connectivity metric (PC) had already demonstrated the importance of including node self-connections in network metrics. As originally defined and used, PC cannot test the importance of self-connections. However, with key terms redefined, PC could be a useful tool in future work.

Node self-connections in network metrics

Zamborain-Mason et al. (Ecol. Lett., 20, 2017, 815-831) state that they have newly proposed network metrics that account for node self-connections. Network metrics incorporating node self-connections, also referred to as intranode (intrapatch) connectivity, were however already proposed before and have been widely used in a variety of conservation planning applications.

Large-Scale Network Analysis of Whole-Brain Resting-State Functional Connectivity in Spinal Cord Injury: A Comparative Study

Network analysis based on graph theory depicts the brain as a complex network that allows inspection of overall brain connectivity pattern and calculation of quantifiable network metrics. To date, large-scale network analysis has not been applied to resting-state functional networks in complete spinal cord injury (SCI) patients. To characterize modular reorganization of whole brain into constituent nodes and compare network metrics between SCI and control subjects, fifteen subjects with chronic complete cervical SCI and 15 neurologically intact controls were scanned. The data were preprocessed followed by parcellation of the brain into 116 regions of interest (ROI). Correlation analysis was performed between every ROI pair to construct connectivity matrices and ROIs were categorized into distinct modules. Subsequently, local efficiency (LE) and global efficiency (GE) network metrics were calculated at incremental cost thresholds. The application of a modularity algorithm organized the whole-brain resting-state functional network of the SCI and the control subjects into nine and seven modules, respectively. The individual modules differed across groups in terms of the number and the composition of constituent nodes. LE demonstrated statistically significant decrease at multiple cost levels in SCI subjects. GE did not differ significantly between the two groups. The demonstration of modular architecture in both groups highlights the applicability of large-scale network analysis in studying complex brain networks. Comparing modules across groups revealed differences in number and membership of constituent nodes, indicating modular reorganization due to neural plasticity.

A novel pattern mining approach for identifying cognitive activity in EEG based functional brain networks

The complex nature of neuronal interactions of the human brain has posed many challenges to the research community. To explore the underlying mechanisms of neuronal activity of cohesive brain regions during different cognitive activities, many innovative mathematical and computational models are required. This paper presents a novel Common Functional Pattern Mining approach to demonstrate the similar patterns of interactions due to common behavior of certain brain regions. The electrode sites of EEG-based functional brain network are modeled as a set of transactions and node-based complex network measures as itemsets. These itemsets are transformed into a graph data structure called Functional Pattern Graph. By mining this Functional Pattern Graph, the common functional patterns due to specific brain functioning can be identified. The empirical analyses show the efficiency of the proposed approach in identifying the extent to which the electrode sites (transactions) are similar during various cognitive load states.

Antagonistic interaction networks are structured independently of latitude and host guild

An increase in species richness with decreasing latitude is a prominent pattern in nature. However, it remains unclear whether there are corresponding latitudinal gradients in the properties of ecological interaction networks. We investigated the structure of 216 quantitative antagonistic networks comprising insect hosts and their parasitoids, drawn from 28 studies from the High Arctic to the tropics. Key metrics of network structure were strongly affected by the size of the interaction matrix (i.e. the total number of interactions documented between individuals) and by the taxonomic diversity of the host taxa involved. After controlling for these sampling effects, quantitative networks showed no consistent structural patterns across latitude and host guilds, suggesting that there may be basic rules for how sets of antagonists interact with resource species. Furthermore, the strong association between network size and structure implies that many apparent spatial and temporal variations in network structure may prove to be artefacts.

Network Sampling and Classification:An Investigation of Network Model Representations

Methods for generating a random sample of networks with desired properties are important tools for the analysis of social, biological, and information networks. Algorithm-based approaches to sampling networks have received a great deal of attention in recent literature. Most of these algorithms are based on simple intuitions that associate the full features of connectivity patterns with specific values of only one or two network metrics. Substantive conclusions are crucially dependent on this association holding true. However, the extent to which this simple intuition holds true is not yet known. In this paper, we examine the association between the connectivity patterns that a network sampling algorithm aims to generate and the connectivity patterns of the generated networks, measured by an existing set of popular network metrics. We find that different network sampling algorithms can yield networks with similar connectivity patterns. We also find that the alternative algorithms for the same connectivity pattern can yield networks with different connectivity patterns. We argue that conclusions based on simulated network studies must focus on the full features of the connectivity patterns of a network instead of on the limited set of network metrics for a specific network type. This fact has important implications for network data analysis: for instance, implications related to the way significance is currently assessed.