Time-varying Network Connectivity Research Articles

Tracking functional network connectivity dynamics in the elderly.

Functional magnetic resonance imaging (fMRI) has shown that aging disturbs healthy brain organization and functional connectivity. However, how this age-induced alteration impacts dynamic brain function interaction has not yet been fully investigated. Dynamic function network connectivity (DFNC) analysis can produce a brain representation based on the time-varying network connectivity changes, which can be further used to study the brain aging mechanism for people at different age stages. This presented investigation examined the dynamic functional connectivity representation and its relationship with brain age for people at an elderly stage as well as in early adulthood. Specifically, the resting-state fMRI data from the University of North Carolina cohort of 34 young adults and 28 elderly participants were fed into a DFNC analysis pipeline. This DFNC pipeline forms an integrated dynamic functional connectivity (FC) analysis framework, which consists of brain functional network parcellation, dynamic FC feature extraction, and FC dynamics examination. The statistical analysis demonstrates that extensive dynamic connection changes in the elderly concerning the transient brain state and the method of functional interaction in the brain. In addition, various machine learning algorithms have been developed to verify the ability of dynamic FC features to distinguish the age stage. The fraction time of DFNC states has the highest performance, which can achieve a classification accuracy of over 88% by a decision tree. The results proved there are dynamic FC alterations in the elderly, and the alteration was found to be correlated with mnemonic discrimination ability and could have an impact on the balance of functional integration and segregation.

Path-dependent dynamics induced by rewiring networks of inertial oscillators.

In networks of coupled oscillators, it is of interest to understand how interaction topology affects synchronization. Many studies have gained key insights into this question by studying the classic Kuramoto oscillator model on static networks. However, new questions arise when the network structure is time varying or when the oscillator system is multistable, the latter of which can occur when an inertial term is added to the Kuramoto model. While the consequences of evolving topology and multistability on collective behavior have been examined separately, real-world systems such as gene regulatory networks and the brain may exhibit these properties simultaneously. It is thus relevant to ask how time-varying network connectivity impacts synchronization in systems that can exhibit multistability. To address this question, we study how the dynamics of coupled Kuramoto oscillators with inertia are affected when the topology of the underlying network changes in time. We show that hysteretic synchronization behavior in networks of coupled inertial oscillators can be driven by changes in connection topology alone. Moreover, we find that certain fixed-density rewiring schemes induce significant changes to the level of global synchrony that remain even after the network returns to its initial configuration, and we show that these changes are robust to a wide range of network perturbations. Our findings highlight that the specific progression of network topology over time, in addition to its initial or final static structure, can play a considerable role in modulating the collective behavior of systems evolving on complex networks.

Decentralized Machine Learning Through Experience-Driven Method in Edge Networks

Data generated at the network edge can be processed locally by leveraging the paradigm of edge computing. To fully utilize the widely distributed data, we concentrate on a wireless edge computing system that conducts model training using decentralized peer-to-peer (P2P) methods. However, there are two major challenges on the way towards efficient P2P model training: limited resources (<i>e.g.</i>, network bandwidth and battery life of mobile devices) and time-varying network connectivity due to device mobility or wireless channel dynamics, which receives less attention in recent years. To address these two challenges, this paper studies the impact of topology construction on the P2P training performance. Specifically, we dynamically construct an efficient P2P topology, where model aggregation occurs at the edge. In a nutshell, we first formulate the topology construction for P2P learning (TCPL) problem with resource constraints as an integer programming problem. Then a learning-driven method is proposed to adaptively construct a topology at each training epoch. We evaluate the performance of our proposed algorithm through extensive simulations and physical platform. Evaluation results show that our method can improve the model training efficiency by about 11&#x0025; with resource constraints, reduce the communication cost by 30&#x0025; and the network traffic consumption by about 60&#x0025; under the same accuracy requirement compared to the benchmarks.

A Scalable Algorithm for Network Localization and Synchronization

The Internet of Things (IoT) will seamlessly integrate a large number of densely deployed heterogeneous devices and will enable new location-aware services. However, fine-grained localization of IoT devices is challenging as their computation and communication resources are typically limited and different devices may have different qualities of internal clocks and different mobility patterns. To address these challenges, we propose a cooperative, scalable, and time-recursive algorithm for network localization and synchronization (NLS). Our algorithm is based on time measurements and supports heterogeneous devices with limited computation and communication resources, time-varying clock and location parameters, arbitrary state-evolution models, and time-varying network connectivity. These attributes make the proposed algorithm attractive for IoT-related applications. The algorithm is furthermore able to incorporate measurements from additional sensors for positioning, navigation, and timing such as receivers for global navigation satellite systems. Based on a factor graph representation of the underlying spatiotemporal Bayesian sequential estimation problem, the algorithm uses belief propagation (BP) for an efficient marginalization of the joint posterior distribution. To account for the nonlinear measurement model and nonlinear state-evolution models while keeping the communication and computation requirements low, we develop an efficient second-order implementation of the BP rules by means of the recently introduced sigma point belief propagation technique. Simulation results demonstrate the high synchronization and localization accuracy as well as the low computational complexity of the proposed algorithm. In particular, in sufficiently dense networks, the proposed algorithm outperforms the state-of-the-art BP-based algorithm for NLS in terms of both estimation accuracy and computational complexity.

Cooperative Simultaneous Localization and Synchronization in Mobile Agent Networks

Cooperative localization in agent networks based on interagent time-of-flight measurements is closely related to synchronization. To leverage this relation, we propose a Bayesian factor graph framework for cooperative simultaneous localization and synchronization (CoSLAS). This framework is suited to mobile agents and time-varying local clock parameters. Building on the CoSLAS factor graph, we develop a distributed (decentralized) belief propagation algorithm for CoSLAS in the practically important case of an affine clock model and asymmetric time stamping. Our algorithm allows for real-time operation and is suitable for a time-varying network connectivity. To achieve high accuracy at reduced complexity and communication cost, the algorithm combines particle implementations with parametric message representations and takes advantage of a conditional independence property. Simulation results demonstrate the good performance of the proposed algorithm in a challenging scenario with time-varying network connectivity.

Higher dimensional analysis shows reduced dynamism of time-varying network connectivity in schizophrenia patients.

Assessments of functional connectivity between brain networks is a fixture of resting state fMRI research. Until very recently most of this work proceeded from an assumption of stationarity in resting state network connectivity. In the last few years however, interest in moving beyond this simplifying assumption has grown considerably. Applying group temporal independent component analysis (tICA) to a set of time-varying functional network connectivity (FNC) matrices derived from a large multi-site fMRI dataset (N=314; 163 healthy, 151 schizophrenia patients), we obtain a set of five basic correlation patterns (component spatial maps (SMs)) from which observed FNCs can be expressed as mutually independent linear combinations, i.e., the coefficient on each SM in the linear combination is maximally independent of the others. We study dynamic properties of network connectivity as they are reflected in this five-dimensional space, and report stark differences in connectivity dynamics between schizophrenia patients and healthy controls. We also find that the most important global differences in FNC dynamism between patient and control groups are replicated when the same dynamical analysis is performed on sets of correlation patterns obtained from either PCA or spatial ICA, giving us additional confidence in the results.