Cross-Scale Energy Coordination in Brain–Body Systems Supports Cognitive Function Across the Lifespan

Cognitive function is commonly framed as neural information processing and formalized in approaches such as active inference, in which metabolism is typically treated as a constraint on network dynamics rather than as an active component of cognition and physiological regulation. This perspective leaves a conceptual gap, in which whole-body processes—particularly peripheral physiology and energy regulation—are not explicitly incorporated as computational contributors to brain homeostasis and cognitive function. Here, we introduce a unified in silico generative framework that integrates whole-brain network dynamics, autonomic oscillations, and a dynamical energy variable into a single dynamical system, enabling explicit bidirectional coupling between neural activity and metabolic regulation. Within this model, Brain–Body coupling is a system-level metric quantifying the alignment between neural metastability and physiological–metabolic fluctuations, capturing coordination not reflected in standard dynamical measures. Across simulated populations, Brain–Body coupling showed limited direct association with synthetic cognitive outcomes but exerted a consistent indirect influence via energy efficiency. Mediation analyses indicate that energy efficiency partially accounts for the relationship between system-level coordination and a metabolism-dependent cognitive proxy. These effects were absent in null simulations and attenuated for energy-independent control outcomes, supporting the specificity of the observed pathway and identifying energy efficiency as a plausible intermediate variable linking brain–body coordination and cognitive performance in the model. To ensure robustness and reduce circularity, we evaluated multiple independent cognitive formulations, including a prediction-based proxy derived from an external dynamical system that does not incorporate metabolic variables. Results were consistent across formulations. Lifespan simulations further revealed a developmental shift in dominant mechanisms, with predictive dynamics more strongly shaping cognitive structure in early life, and energy-related influences becoming increasingly prominent with age. Collectively, this work provides a unified generative framework linking neural dynamics, physiology, and brain metabolism, and demonstrates that large-scale neural coordination may contribute not only to computation but also to the regulation of metabolic efficiency in coupled brain–body systems. Although based on simulations, the model generates empirically testable hypotheses regarding how energy regulation shapes brain–body coordination and cognitive variability across the lifespan.