Beyond Networks: Toward Adaptive Models of Biological Complexity

Network models have transformed our understanding of complex systems across biology, technology,and society, proving valuable in neuroscience. However, modeling biological complexityposes specific challenges, calling for expansions of traditional network frameworks. This paperexplores constructive ways to enhance models, highlighting opportunities such as incorporatingtime-varying connections, adaptive topologies, and multilayer structures to better represent thetemporal dynamics and multi-level interactions characteristic of biological systems. Additionally,it addresses deeper conceptual challenges, notably the substantial context dependence,open-endedness, and history sensitivity often observed in biology. By introducing conceptssuch as Kauffman’s "adjacent possible," the discussion emphasizes how biological state spacesthemselves may dynamically evolve, suggesting the need for modeling strategies beyond staticor pre-specified assumptions. Rather than undermining network science, these considerationshighlight areas where traditional formalisms can fruitfully adapt and grow, ultimately deepeningtheir explanatory power. The paper advocates integrating data-driven approaches that dynamicallyinfer system properties from empirical observations, balancing modeling generality withbiological specificity. Overall, this synthesis provides an assessment of both the strengths ofnetwork science and the challenges it faces, proposing constructive avenues for methodologicaland conceptual innovation that advance our ability to capture the nuanced complexity inherent inbiological phenomena.