This final analysis provides a robustness check, using a behaviorally-derived G factor as the model’s outcome, to assess the consistency of the generative model across parameters and outcome measures Brain–Body–Energy Generative Model This repository contains a fully parallelized in-silico simulation framework designed to explore how brain–body–energy interactions shape cognition across lifespan-like cohorts. The model is inspired by systems-level theories suggesting that cognitive outcomes emerge not only from pathology, but from how metabolic stability and neural dynamics interact across scales. Scientific Motivation Traditional approaches to cognitive vulnerability often focus on molecular or pathological causes. Here, we take a different/complemetary perspective: Cognition is an emergent property of the whole brain–body system. Large-scale brain dynamics are constrained by energy balance and autonomic signals. Cognitive performance depends on the alignment between neural metastability and peripheral energy fluctuations, captured by what we call the Energy Coupling By simulating synthetic cohorts across a wide age range, we can explore how coupling strength influences cognition, revealing patterns like the inverted-U trajectory, where optimal cognition emerges at intermediate coupling and declines at extremes. In short, this framework allows researchers to test mechanistic hypotheses about energy, brain dynamics, and cognition, without relying on real-world pathology data. What This Model Does Simulates brain networks Small-world networks with tunable size and connectivity. Network dynamics modeled using graph Laplacians. Models body oscillators Cardiac and respiratory rhythms for each synthetic subject. These peripheral signals interact bidirectionally with brain activity. Simulates brain–energy dynamics Neural activity modeled via Stuart-Landau oscillators. Metabolic energy is dynamically produced and consumed based on brain and body signals. Coupling parameters (K_BODY, K_ENERGY) control the influence of body and energy on the brain. Computes metrics of system behavior Metastability: how synchrony fluctuates across brain regions. Brain–Heart Coherence: phase alignment between cardiac and neural signals. Forward Modeling: predictive variability of neural power. Energy Stability and Efficiency: variability and mean of metabolic energy. Synthetic Cognition: linear combination of energy, body signals, and stochastic noise. Fully parallelized Uses all available CPU cores to simulate large cohorts quickly. Supports multiple network sizes, random seeds, and hundreds of subjects per network. Outputs The simulation produces a comprehensive dataset of synthetic subjects: BrainBodyEnergy_Publication/dataset_full.csv Columns include: network ID, subject ID, age, coupling parameters, brain–body metrics, energy metrics, and synthetic cognition. A NullModel column is included as a control. Recommended Analyses Partial correlations between predictors (BrainHeartCoherence, ForwardModeling, NullModel). Robust regression predicting synthetic cognition. Mediation analysis: EnergyEfficiency as a mediator between brain–body metrics and cognition. Lifespan simulations: explore inverted-U trajectories of cognition versus coupling strength. Running the Simulation git clone cd BrainBodyEnergy_Publication python brain_body_energy_parallel.py The script automatically detects your CPU cores and runs simulations in parallel. Results are saved as .csv files for analysis. Dependencies Python ≥ 3.10 numpy pandas scipy scikit-learn statsmodels joblib matplotlib seaborn Notes Fully in-silico — no human or animal data is required. Designed for exploring mechanistic hypotheses linking energy, brain dynamics, and cognition. Modular design allows easy extensions to new network topologies, coupling parameters, and cognitive metrics. If you want, I can also write a short, one-paragraph “lay summary” for the README front page — something a non-specialist could read in a few seconds and immediately understand the purpose and value of this model. Do you want me to do that? add the reference please Absolutely! Here’s your fluid, human-readable README description with the reference added at the end: Brain–Body–Energy Generative Model This repository contains a fully parallelized in-silico simulation framework designed to explore how brain–body–energy interactions shape cognition across lifespan-like cohorts. The model is inspired by systems-level theories suggesting that cognitive outcomes emerge not only from pathology, but from how metabolic stability and neural dynamics interact across scales. Scientific Motivation Traditional approaches to cognitive vulnerability often focus on molecular or pathological causes. Here, we take a different perspective: Cognition is an emergent property of the whole brain–body system. Large-scale brain dynamics are constrained by energy balance and autonomic signals. Cognitive performance depends on the alignment between neural metastability and peripheral energy fluctuations. By simulating synthetic cohorts across a wide age range, we can explore how coupling strength influences cognition, revealing patterns like the inverted-U trajectory, where optimal cognition emerges at intermediate coupling and declines at extremes. In short, this framework allows researchers to test mechanistic hypotheses about energy, brain dynamics, and cognition, without relying on real-world pathology data. What This Model Does Simulates brain networks Small-world networks with tunable size and connectivity. Network dynamics modeled using graph Laplacians. Models body oscillators Cardiac and respiratory rhythms for each synthetic subject. These peripheral signals interact bidirectionally with brain activity. Simulates brain–energy dynamics Neural activity modeled via Stuart-Landau oscillators. Metabolic energy is dynamically produced and consumed based on brain and body signals. Coupling parameters (K_BODY, K_ENERGY) control the influence of body and energy on the brain. Computes metrics of system behavior Metastability: how synchrony fluctuates across brain regions. Brain–Heart Coherence: phase alignment between cardiac and neural signals. Forward Modeling: predictive variability of neural power. Energy Stability and Efficiency: variability and mean of metabolic energy. Synthetic Cognition: linear combination of energy, body signals, and stochastic noise. Fully parallelized Uses all available CPU cores to simulate large cohorts quickly. Supports multiple network sizes, random seeds, and hundreds of subjects per network. Outputs The simulation produces a comprehensive dataset of synthetic subjects: BrainBodyEnergy_Publication/dataset_full.csv Columns include: network ID, subject ID, age, coupling parameters, brain–body metrics, energy metrics, and synthetic cognition. A NullModel column is included as a control. Recommended Analyses Partial correlations between predictors (BrainHeartCoherence, ForwardModeling, NullModel). Robust regression predicting synthetic cognition. Mediation analysis: EnergyEfficiency as a mediator between brain–body metrics and cognition. Lifespan simulations: explore inverted-U trajectories of cognition versus coupling strength. Running the Simulation git clone cd BrainBodyEnergy_Publication python brain_body_energy_parallel.py The script automatically detects your CPU cores and runs simulations in parallel. Results are saved as .csv files for analysis. Dependencies Python ≥ 3.10 numpy pandas scipy scikit-learn statsmodels joblib matplotlib seaborn Notes Fully in-silico — no human or animal data is required. Designed for exploring mechanistic hypotheses linking energy, brain dynamics, and cognition. Modular design allows easy extensions to new network topologies, coupling parameters, and cognitive metrics. References Lombardo, D.M. (2026). Energy Coupling in Brain–Body Systems: An In-Silico Approach to Lifespan Cognitive and Computational Vulnerability.