Artificial Intelligence and Electrical & Electronics Engineering: AIEEE Open Access
Multiscale Dynamic Network Integration: From Molecular Fluctuations to Organismal Phenotypes: A Conceptual Framework for Bridging Temporal and Spatial Scales in Biological Networks
Abstract
Erwin L. Rimban
Biological function unfolds across intertwined temporal and spatial scales—from millisecond molecular fluctuations to organismal adaptations over years—yet most modeling traditions succeed only within single strata of this hierarchy. Recent work in systems biology has underscored the organizing role of network structure and dynamic motifs, but less attention has been paid to how causality and information transit across scales [1-3]. This paper advances a conceptual framework—Theory of Multiscale Network Integration (TMNI)—that treats cross-scale interactions as structured, dynamical translations among networked subsystems. We first motivate the scale-integration problem and propose a taxonomy of scale-bridging phenomena: emergence, downward causation, scale-specific feedback loops, and information bottlenecks/hubs. We then critically review existing formalisms—coarse-graining, hierarchical statistical models, and multiscale information metrics—highlighting their contributions and limitations. TMNI rests on four principles: primacy of dynamics, network motifs as units of cross-scale communication, context-dependent scaling rules, and abstract translation layers mediating biochemical and physiological descriptions. We outline mathematical directions (category theory, sheaf theory), computational needs (integrated platforms), and validation strategies (in-silico instantiations). Overall, TMNI aims to reorient multiscale biology from descriptive aggregation toward principled, predictive integration of causal dynamics.