Source-linked AI summary
Structure and dynamics of core-periphery networks
Peter Csermely, Andras London, Ling-Yun Wu, Brian Uzzi
TL;DR
The paper addresses how core/periphery structures and related network organizations should be characterized across complex systems. It reviews traditional cores, rich-clubs, nested, bow-tie, and onion networks, comparing their structures and dynamics. The review highlights robustness through degenerate pathways and flow re-channelling, while identifying unresolved conditions governing core formation and transformation.
Problem
Core/periphery networks are important across complex systems, but their structure, dynamics, function, and distinctions from related organizations remain incompletely characterized.
Method
The paper reviews traditional core/periphery networks, rich-clubs, nested networks, bow-tie structures, and onion networks, comparing global and local cores and related topologies.
Results
Core/periphery organization is associated with robustness through degenerate pathways, cooperation, flow re-channelling, coordinated responses, reduced fluctuations, and slow core evolution.
Takeaways & Limitations
Core/periphery structures integrate, stabilize, and support evolvability in diverse complex systems, including social, ecological, neuronal, industrial, and Internet networks.
Takeaways & Limitations
The environmental and network conditions governing the number, size, and abrupt transformation of network cores remain unclear.
Abstract
from arXiv · showhide
Recent studies uncovered important core/periphery network structures characterizing complex sets of cooperative and competitive interactions between network nodes, be they proteins, cells, species or humans. Better characterization of the structure, dynamics and function of core/periphery networks is a key step of our understanding cellular functions, species adaptation, social and market changes. Here we summarize the current knowledge of the structure and dynamics of "traditional" core/periphery networks, rich-clubs, nested, bow-tie and onion networks. Comparing core/periphery structures with network modules, we discriminate between global and local cores. The core/periphery network organization lies in the middle of several extreme properties, such as random/condensed structures, clique/star configurations, network symmetry/asymmetry, network assortativity/disassortativity, as well as network hierarchy/anti-hierarchy. These properties of high complexity together with the large degeneracy of core pathways ensuring cooperation and providing multiple options of network flow re-channelling greatly contribute to the high robustness of complex systems. Core processes enable a coordinated response to various stimuli, decrease noise, and evolve slowly. The integrative function of network cores is an important step in the development of a large variety of complex organisms and organizations. In addition to these important features and several decades of research interest, studies on core/periphery networks still have a number of unexplored areas.
1. Introduction
Core/periphery networks combine a densely connected central set with a sparse periphery linked to it, and arise across many complex systems. Their study requires distinguishing core structures from modules and accounting for robustness, null-model choice, and network representation.
- A network core is typically central and densely connected, whereas the periphery is sparse, usually non-central, and linked to the core.
- Core structures can support robustness and evolvability because densely intertwined pathways can substitute for or support one another.
- Core/periphery organization differs from network modules and can be classified among four broad topological classes using spectral scaling.
- Core/periphery structures have been identified in biological, ecological, social, economic, engineered, and information networks.
- Null-model selection and interpretation are essential but difficult when defining and comparing network-core properties.
- Edge-cores are especially relevant in weighted or directed networks but were excluded from detailed discussion because most edge-skeletons are not densely interconnected.
2. Definitions and structural properties of core/periphery networks
The review defines core/periphery organization as dense, central core structure linked to a sparse periphery, and surveys related forms including rich-clubs, nested, bow-tie, and onion networks. It also emphasizes algorithmic detection, null-model assessment, and structural distinctions across these network types.
- The review covers traditional core/periphery networks alongside rich-clubs, nested structures, bow-tie organization, and onion networks.These structures are treated as related forms of network organization with distinct structural or dynamical properties.
- Borgatti–Everett models an ideal network with a fully linked core, a periphery connected to the core, and no links among peripheral nodes.The formal definition compares the observed adjacency matrix with an ideal core/periphery matrix.
- Core detection is posed as finding the core-membership vector that maximizes correlation between the observed adjacency matrix and the ideal core/periphery matrix.The ideal matrix is constructed from the binary core-membership vector δ.
- Null models are essential because random graphs, including dense or power-law networks, can contain apparently dense cores or increasing rich-club coefficients.Appropriate randomized benchmarks are therefore needed to assess whether detected core or rich-club structure exceeds chance expectations.
- Rich-club behavior can reverse across connectivity layers, while hub dissociation can maximize network observability and node failures can produce tricritical behavior.These findings show that rich-club effects depend on network layer and interdependence.
- Bow-tie networks contain source-driven fan-in and sink-directed fan-out surrounding a highly intertwined core, whereas onion networks are rewired structures designed for robustness.BowTieBuilder quantifies bow-tie-ness by searching probable source-to-sink pathways; no real-world onion examples had yet been found.
3. Dynamics of core/periphery networks
Core/periphery structures can emerge or transform as resources, stress, cooperation, and environmental conditions change. Core size creates a trade-off between controllability and adaptability, while reduced cores can introduce signalling bottlenecks and cores can enhance robustness through redundant pathways.
- Core/periphery structures may develop under low resources or increased environmental stress, alongside more condensed structures or separated network modules.
- Flow-type networks more often develop characteristic core/periphery structures than association-type networks, partly because edge-length costs affect their organization.
- Cooperation can segregate a social-network core, while nested ecological networks reduce effective inter-specific competition.
- Resource scarcity or severe stress may shrink a core to a single hub or transform a core/periphery structure into a chain, but governing conditions remain unclear.
- Smaller cores may tighten controllability while reducing flexibility and adaptability; larger cores may increase behavioral plasticity.
- Core reduction can create signalling bottlenecks, reduce signalling capacity under noise, and diminish responsiveness to external stimuli.
- Core development increases robustness and stability because dense core connections support degenerate processes, cooperation, and multiple flow-rechannelling options.
4. Function of core/periphery network structures in different types of real world networks
Core/periphery organization supports conserved, coordinated, and robust functions across molecular, cellular, social, economic, and engineered networks. Its specific structural roles include routing flows, concentrating control, organizing cooperation, and preserving network stability under environmental or structural change.
- Molecular networks: Protein interactomes place conserved, often essential general-function proteins in the core and organ-specific proteins in the periphery.Peripheral proteins tend to localize toward the cell surface, while low-connectivity peripheral positions are associated with preferential age-related DNA methylation changes.
- Molecular networks: Metabolic bow-tie cores contain conserved reactions, while core–fan-in and core–fan-out connections identify vulnerable links relevant to drug targeting.Core size varies across organisms, with larger cores in organisms living under constant environmental conditions.
- Molecular networks: Signalling and gene-regulatory networks route inputs through core transcription factors toward regulated genes, whose master regulators are vital and nearly continuously active.In signalling networks, receptors typically form the fan-in, transcription factors the core, and induced genes and regulatory microRNAs the fan-out.
- Cellular networks: Rich-club brain organization provides a synchronized, regulatable, high-capacity communication backbone, while learning separates a relatively stiff sensorimotor core from a flexible association-region periphery.The brain backbone carries high communication cost, estimated at 40% of total, and task practice changes the core–periphery separation.
- Ecological networks: Nested structures can enhance robustness and resilience in mutualistic systems, but node contributions are heterogeneous and removing strong contributors can reduce persistence.Nestedness may emerge from abundance optimization, while the rarest species’ abundance is directly linked to community resilience.
- Social, economic, and engineered networks: Social, economic, and engineered networks show core structures that organize cooperation, concentrate control, preserve nestedness, or stabilize systems despite shocks and rapid development.Examples include cooperative hubs excluding defectors, persistent economic nestedness, shock-related increases in artistic-network nestedness, and a stable Internet core.
5. Conclusions and perspectives
The review identifies unresolved questions about defining, distinguishing, and measuring core/periphery structures, while highlighting their robustness-related functions and broad relevance. It concludes that substantial opportunities remain for clarifying their structure, dynamics, and function.
- Open questions: Distinguishing network cores from modules and global cores from local cores still requires appropriate mathematical formalism and further work.The review points to spectral scaling and node/edge dynamics as possible routes for deeper discrimination.
- Conclusions: Core/periphery structures combine diverse network properties and rich pathway degeneracy, contributing to robustness and multiple options for flow re-channelling.Their cooperation-supporting pathways can substitute for or support one another when needed.
- Conclusions: Core processes support coordinated responses to stimuli, fluctuate less than peripheral processes, and evolve more slowly under stronger constraints.The integrative function of network cores is described as important across many complex systems.
- Conclusions: Cooperation links integration, stabilization, and evolvability across examples including social dilemma games, ecological networks, neuronal networks, industrial networks, and the Internet.The review presents these cases as evidence of core/periphery structures in fast-growing, fast-changing systems.
- Open questions: Open questions concern general definitions, useful core measures, onion-network counterparts in real systems, signalling functions, spatial embedding, and hub-hub association or avoidance.The review also identifies unresolved questions about protein structure networks and dynamic changes during allosteric signalling.
- Perspectives: Despite decades of research, core/periphery networks retain substantial scope for further discoveries.The authors explicitly frame the field as promising and unfinished.
Funding
The authors report support from national research foundations, Northwestern University’s Institute on Complex Systems, and the Army Research Laboratory.
- Funding: The authors’ groups received support from Hungarian and Chinese national science foundations, Northwestern University’s Institute on Complex Systems, and the Army Research Laboratory.The funding statement lists specific grants and a cooperative agreement, while noting that the views and conclusions are those of the authors.
k-truss or k-dense subgraph
The supplied passages define k-core structures and provide illustrative comparisons of core/periphery network forms. They do not provide a definition of k-truss or k-dense subgraphs.
- Definitions: A k-core is a maximal connected subgraph whose elements connect to at least k other elements within that subgraph.It can alternatively be represented as the union of all k-shells with indices at least k.
- Definitions: A k-shell consists of nodes and belonging links removed consecutively under a degree threshold of ≤ k.The passage presents this node-removal formulation as an alternative basis for defining k-cores.
- Scope: The supplied figure passages contrast bow-tie, rich-club, nested, and onion structures rather than defining k-truss or k-dense subgraphs.Figure 1 presents general core/periphery features, while Figure 2 illustrates several network topologies.