Centrality Measures in Network Analysis

Centrality measures quantify the structural importance of nodes in a network by assigning each node a numerical score that reflects its position and potential influence. Classic indicators include degree centrality, which counts direct connections; betweenness centrality, which identifies nodes acting as bridges on shortest paths; closeness centrality, which evaluates access to all other nodes via minimal routes; and eigenvector centrality, which rewards connections to other well-connected nodes. These metrics have underpinned insights into social cohesion, information diffusion, infrastructure resilience and molecular interactions. Recent theoretical work has unified diverse centrality definitions under a single framework derived from the adjacency matrix, enabling multi-component measures that capture complementary aspects of network influence. Advances in computation and data availability have expanded applications to dynamic and weighted networks, percolation processes and strategic settings in which nodes may adapt to evade detection. Comparative studies of centrality correlations across hundreds of real-world networks have revealed how topology—modularity, density and degree distribution—modulates the behaviour and interpretability of different measures. The global significance of centrality analysis spans epidemiology, urban planning, cybersecurity and beyond, offering a common language for identifying critical nodes, forecasting systemic risk and guiding interventions in complex systems.

Research from Nature Portfolio

Recent experimental work has uncovered a genetic basis for betweenness centrality in a model organism, demonstrating that allelic variation in the dokb gene modulates an individual’s role as an information conduit within a social network. By swapping alleles between strains and tracking behavioural patterns, researchers linked molecular pathways to network cohesion, opening avenues to explore conserved mechanisms of social influence. A complementary theoretical study proposed a general framework in which centrality measures naturally emerge from the spectral decomposition of the adjacency matrix. This approach unifies degree, eigenvector and hub-authority metrics, identifies intrinsic limitations of single-component scores and introduces multi-dimensional centralities that outperform standard measures across diverse network topologies.

Research from all publishers

A game-theoretic model has characterised the interaction between an analyst seeking to identify high-centrality nodes and an evader attempting to minimise detection. Framed as a Stackelberg game, the work establishes NP-completeness for optimal edge modifications, offers approximation algorithms for degree, closeness and betweenness centralities, and formulates Mixed Integer Linear Programming solutions for equilibrium strategies. An interdisciplinary study connected degree centrality to expected-utility theory, axiomatizing which centrality measures correspond to risk-neutral preferences in cooperative games. The authors showed that only degree-based measures satisfy neutrality to ordinary risk while accommodating externalities via neighbour counts, thereby providing a utility foundation for a class of Shapley-like centralities. A large-scale empirical analysis of over two hundred networks examined correlations among seventeen centrality metrics, demonstrating that network modularity and density drive the strength of pairwise associations. Data-driven clustering of nodes by centrality profiles revealed consistent roles—core, bridge and peripheral—which enhances the interpretability of nodal importance across disciplines.

Centrality Measures in Network Analysis publication trend

The graph below shows the total number of articles in centrality measures in network analysis across all publications each year (not limited to Nature Index journals).

Technical terms

Node: Fundamental unit of a network representing an individual entity or actor.

Degree centrality: Count of immediate connections a node has, indicating local influence.

Betweenness centrality: Measure of how often a node lies on shortest paths between other node pairs.

Eigenvector centrality: Score reflecting a node’s influence based on the importance of its neighbours.

Closeness centrality: Inverse of the average shortest path distance from a node to all others, indicating overall accessibility.

Adjacency matrix: Matrix encoding the presence or weight of edges between every pair of nodes in a network.

Stackelberg game: Sequential game-theoretic model with a leader and a follower making strategic decisions.

NP-complete: Class of decision problems for which no polynomial-time solution is known and to which any NP problem can be reduced.

References

1. The gene “degrees of kevin bacon” (dokb) regulates a social network behaviour in Drosophila melanogaster. Nature Communications (2024).
2. Hiding From Centrality Measures: A Stackelberg Game Perspective. IEEE Transactions on Knowledge and Data Engineering (2023).
3. Degree centrality, von Neumann–Morgenstern expected utility and externalities in networks. European Journal of Operational Research (2024).
4. Consistency and differences between centrality measures across distinct classes of networks. PLOS ONE (2019).
5. A change of perspective in network centrality. Scientific Reports (2018).