| name | competition-stability-ei-circuits |
| description | Game-theoretic energetic framework for excitatory-inhibitory neural circuits - competition, stability, and functionality in asymmetric networks |
| category | neuroscience |
| created | 2026-06-04T00:00:00.000Z |
| arxiv_id | 2512.05252 |
| authors | Simone Betteti, William Retnaraj, Alexander Davydov, Jorge Cortés, Francesco Bullo |
| status | available |
| dependencies | [] |
| activation_keywords | excitatory-inhibitory networks, game theory, energy-based models, asymmetric networks, neural stability, Wilson-Cowan, lateral inhibition, cortical columns, contrast enhancement |
Competition, Stability, and Functionality in E-I Neural Circuits
Overview
Methodology from arXiv:2512.05252 (v2, revised 3 Jun 2026) that extends energetic frameworks to asymmetric excitatory-inhibitory (E-I) networks using game-theoretic structure.
Core Innovation: Each neuron is modeled as an agent minimizing its own energy, enabling systematic analysis of asymmetric neural systems where classical energy landscape theory fails.
Key Contributions
1. Game-Energetic Framework
- Asymmetric Networks: Extends energetic framework beyond symmetric weight matrices
- Game Theory Structure: Neurons as agents seeking energy minimization
- Biological Realism: Accounts for E/I constraints absent in classical models
2. Stability Principles
- Network Theory Integration: Rigorous stability principles from network control
- Activity Regulation: Study regulation and balancing of neural activity
- Dynamic Stability: Systematic engineering of stable architectures
3. Cortical Functionality
- Wilson-Cowan Model: Revisited with game-energetic interpretation
- Lateral Inhibition: Microcircuit analysis as contrast enhancer
- Cortical Columns: Hierarchical E/I interplay for subtle difference sharpening
Technical Details
Problem Context
Energy-based models rely on symmetry in synaptic matrices - excluding biologically realistic E-I networks. When symmetry relaxes, global energy landscape fails, leaving asymmetric dynamics conceptually unanchored.
Solution Mechanism
Game-Theoretic Interpretation:
- Each neuron = agent minimizing local energy
- Competition emerges from E/I constraints
- Stability from network-theoretic principles
Mathematical Framework
- Asymmetric Firing Rate Networks: Extended energetic framework
- Network Stability Principles: Control theory integration
- Game Theory: Agent-based energy minimization
Key Properties
- Local Energy Minimization: Per-neuron optimization
- Competitive Dynamics: E/I induced competition
- Stable Equilibria: Network-theoretic stability guarantees
