neural-avalanche-utac

GenesisAeon Package 20 (P20) — Brain Criticality & Neuronal Avalanches as UTAC System

Neuronal avalanches at criticality modelled as UTAC system.

Key result: Γ_brain ≈ 0.251, a rescaling of the η = 50% homeostatic setpoint through the shared GenesisAeon σ=2.2 default. Its numeric match to amoc-utac's Γ_AMOC is a verified construction artifact of that shared constant (both packages use the identical formula and coincidentally equal η), not an independent cross-domain physics finding — see constants.py's GAMMA_BRAIN docstring for the full derivation.

Installation

pip install neural-avalanche-utac

For local development:

pip install -e ".[dev]"

Quickstart

neural-utac run --duration 3600
neural-utac criticality-check
neural-utac gamma-universality

Integration in genesis-os

from genesis_os import GenesisOS
os = GenesisOS()
neural = os.load_package(20)
results = neural.run_cycle(duration_seconds=3600)

Benchmark

Validated against Hengen & Shew (2025).

On the "50% setpoint → Γ ≈ 0.251" claim

This is NOT a falsifiable prediction as stated: for ANY system with an η = 50% efficiency/homeostatic setpoint, arctanh(0.50)/σ gives exactly this value at σ=2.2 by pure algebra — the formula guarantees the match regardless of the underlying domain's physics. A genuine test of cross-domain universality would need σ independently derived per domain against real dose-response data, which has not been done here or in any other GenesisAeon UTAC package sharing this default. See constants.py's GAMMA_BRAIN docstring for the full reasoning.

Role in the GenesisAeon Ecosystem

neural-avalanche-utac is GenesisAeon Package P20 (domain: neuroscience / cortical criticality). It models neuronal avalanches at criticality as a UTAC dynamical system, giving Γ_brain ≈ 0.251 — a rescaling of the η=50% homeostatic setpoint through the shared σ=2.2 default (see above for why this is not independent cross-domain evidence).

Citation

See CITATION.cff for the package citation and the GenesisAeon whitepaper DOI badge above.

License

MIT