Title : Emergent neural phenotypes from local developmental optimization: A mechanistic framework for brain organization and neurodiversity
Abstract:
A central question in neuroscience is how complex, structured brain architectures emerge from local cellular interactions during development. Here, we demonstrate that local developmental optimization alone is sufficient to generate realistic brain organization and stable neurodivergent trajectories, without requiring global supervision or centralized control. We present a mechanistic, simulation-based framework in which individual neurons act as local agents balancing competing biological constraints, including information-processing efficiency, wiring economy, and metabolic cost.
Neuronal activity, synaptic plasticity, adaptive tradeoff regulation, and spatial growth are jointly governed by a unified local energy principle. Stochastic and entropy-regularized mechanisms promote developmental exploration, while slow adaptive dynamics stabilize emerging network structure. This framework produces robust developmental trajectories that naturally transition through phases of growth, modular differentiation, synaptic pruning, and stabilization.
Across a wide range of conditions, the model reproduces key features of cortical organization, including hierarchical modularity (modularity index Q ~0.45–0.6), small-world connectivity (small-world index σ ~1.5–1.7), wiring-cost scaling laws, and characteristic synaptic pruning curves. These emergent properties arise directly from local interactions, demonstrating that large-scale brain architecture can be explained as a collective outcome of simple neuronal optimization rules.
By systematically varying developmental tradeoffs, stochasticity, and adaptive learning rates, the model generates a diverse ensemble of stable neural phenotypes, forming distinct developmental attractors rather than pathological deviations. These trajectories capture a continuum of structural and functional brain organizations, offering a mechanistic interpretation of neurodiversity as adaptive specialization under competing biological constraints. Divergence across trajectories is quantified using modularity, hierarchy, network efficiency, and rewiring dynamics, revealing structured variation rather than disorder.
The framework is computationally scalable and dimensionless, allowing systematic exploration of developmental regimes across network sizes and parameter spaces. Logging of phase-specific network statistics, motif persistence, and efficiency–cost tradeoffs enables detailed analysis of robustness, adaptability, and sensitivity to perturbations.
Together, these results provide a principled link between local cellular processes and macroscopic brain organization, showing how structured neural architectures and functional diversity emerge naturally from developmental optimization. More broadly, this work suggests that brain diversity reflects adaptive solutions to fundamental developmental tradeoffs, providing a unifying mechanistic account of both typical and neurodivergent neurodevelopment.
