Energy-Driven Structure in Coupled Brain–Body Dynamics

Technical Summary: Cross-Scale Energy Coordination in Brain–Body Systems

Problem Statement
Neuropsychiatric and neurodegenerative disorders are traditionally framed as downstream consequences of molecular pathology cascades leading to network dysfunction. However, the limited success of pathology-focused interventions suggests a need for complementary systems-level explanations that account for early vulnerability before irreversible structural damage occurs. Current models often assume behavioral optimization or prediction-error minimization as the central organizing principle of brain function. This study proposes an alternative perspective: that large-scale brain dynamics are primarily organized to regulate energy across distributed brain–body systems, and that cognitive vulnerability emerges when this cross-system energy regulation becomes inefficient.

Methodology
The author developed a generative in silico framework to investigate the interaction between large-scale neural dynamics, peripheral physiology, and metabolic energy regulation. The study utilized the following methodological components:

• Model Architecture: A whole-brain model combining networks of coupled nonlinear Stuart–Landau oscillators with a simulated cardiac signal and a dynamical energy variable.
• Network Construction: Structural connectivity was modeled using multiple canonical topologies (Small-world, Random, and Lattice networks) with node counts of 40, 80, and 120 to ensure results were not architecture-dependent.
• Dynamics:
  • Neural: Nodes evolved via intrinsic oscillatory dynamics, diffusive coupling (via graph Laplacian), and multiplicative modulation from cardiac input and metabolic energy.
  • Peripheral: A stochastic cardiac oscillator provided a time-varying external input to neural dynamics.
  • Energy: Metabolic energy was modeled as a dynamical variable balancing production (driven by the cardiac signal), consumption (scaling with neural power), and decay. This established bidirectional coupling where neural activation consumes energy, and available energy modulates neural dynamics.
• Lifespan Parameterization: Key parameters (coupling strength, neural noise, metabolic decay) were modulated as functions of age to simulate developmental and degenerative trajectories without imposing discrete groupings.
• Metrics:
  • Brain–Heart Coherence (BHC): A metric quantifying the alignment between neural metastability and autonomic–metabolic variability (phase coherence between global neural signal and cardiac rhythm).
  • Energy–Entropy Coupling Index (EECI): A dynamic systems metric quantifying the alignment between network metastability and autonomic–metabolic variability.
  • Synthetic Cognition: A proxy variable combining global coupling strength with energy statistics (mean energy and variability).
• Analysis: Statistical associations were evaluated using Pearson correlation and permutation testing. Mediation analyses (bootstrap resampling) were performed to test if energy efficiency mediated the relationship between system-level coordination and cognitive outcomes. Longitudinal trajectories were analyzed across 5-year age bins (20–85 years) with FDR correction.

Key Contributions

1. Introduction of Brain–Heart Coherence (BHC): A novel metric quantifying the alignment between neural metastability and autonomic–metabolic variability, formalizing neurovisceral integration within a dynamical systems framework.
2. Systems-Level Hypothesis: Proposes that cognitive performance depends on the efficiency of cross-scale energetic coordination rather than pathology burden alone. It suggests that cognitive vulnerability arises when the system deviates from an energy-constrained metastable optimum.
3. Differentiation of Mechanisms: Demonstrates that energy–entropy coupling provides explanatory value distinct from predictive-processing measures (Active Inference) and pathology-centric models.
4. Lifespan Trajectories: Identifies emergent trajectories where optimal cognition occurs within specific energy-constrained metastable regimes, showing distinct temporal peaks for different models (Active Inference peaking in early adulthood; Brain–Body coupling peaking in mid-to-late adulthood).

Results

• Predictive Validity: The Active Inference model showed a moderate association with "Cognitive Energy" ($R^2 \approx 0.089$), whereas Brain–Body coupling and Null models showed negligible effects ($R^2 < 0.001$). However, the predictors were statistically independent, capturing non-overlapping variability.
• Mediation by Energy Efficiency:
  • For the Brain–Body model, the association with cognitive energy was primarily expressed through a mediated pathway via energy efficiency (negative indirect effect, positive direct effect).
  • For the Active Inference model, both indirect and direct effects were large and positive.
  • The Null model showed negligible mediation effects.
• Lifespan Dynamics:
  • Energy efficiency showed a monotonic decline across the adult lifespan.
  • The Active Inference model displayed its strongest mediated effect in early adulthood (peak at 22.5 years).
  • The Brain–Body model reached its maximal effect later in life (peak at 57.5 years).
  • These trajectories were robust across cross-validation folds and network topologies.
• Construct Specificity: The findings demonstrated that the observed relationships were specific to the brain–body–energy system and not artifacts of random variability or circular definitions.

Significance and Claims
The paper claims to provide a mechanistically explicit and empirically testable model linking neural dynamics, metabolism, and peripheral physiology. It argues that cognitive vulnerability should be viewed as a systems property of brain–body energetic alignment across the lifespan.

• Theoretical Shift: The work motivates a shift from viewing cognitive decline solely as a cumulative burden of pathology or "wear and tear" (allostatic load) to viewing it as a loss of energetic efficiency and a deviation from an energy-constrained metastable optimum.
• Developmental Relevance: The model suggests that the developmental window where major psychiatric disorders emerge (adolescence and early adulthood) coincides with specific energetic trade-offs and shifts in coupling strength, potentially explaining vulnerability before structural damage occurs.
• Limitations and Scope: The author explicitly states the study is entirely in silico with simplified peripheral and metabolic representations. They do not claim to replace molecular models but to offer a complementary hypothesis. The small $R^2$ values are attributed to the independent design of the synthetic cognitive variable, and the author acknowledges that future work requires empirical validation with multimodal neuroimaging and physiological recordings.