Astrocyte-Aware Regulatory Stack (Control Plane Mechanism)

Claim Type: mechanism_hypothesis
Scope: Slow regulatory field mediating control-plane signals and precision routing
Depends On: ARC-005 (control plane), INV-008 (precision routing), ARC-004 (L-space)
Status: candidate
Claim ID: MECH-001

Regulatory Stack Model (Excerpt)

This document introduces the regulatory stack framing: a layered architecture in which monoamines are reinterpreted as meta-regulatory broadcast signals rather than direct control knobs.

The stack (top to bottom)

┌─────────────────────────────────────────────────────────────────┐
│  Monoamines (Dopamine, Serotonin, Noradrenaline, etc.)         │
│  ───────────────────────────────────────────────────────────    │
│  Broadcast meta-regulatory signals (tags / requests)            │
│  • "commit to this" (DA-like)                                   │
│  • "hold options open" (5-HT-like)                              │
│  • "reset / increase gain" (NA-like)                            │
│  • "focus sensory processing" (ACh-like)                        │
└───────────────────────┬─────────────────────────────────────────┘
                        ↓
┌─────────────────────────────────────────────────────────────────┐
│  Astrocytic Regulatory Field                                    │
│  ───────────────────────────────────────────────────────────    │
│  Slow, spatially resolved, history-dependent control substrate  │
│  • Integrates monoaminergic tone over seconds-to-minutes        │
│  • Modulated by local activity, metabolic state, past history   │
│  • Coupled across space (gap junction syncytium)                │
│  • State: R(x,t) — regulatory field over L-space positions      │
└───────────────────────┬─────────────────────────────────────────┘
                        ↓
┌─────────────────────────────────────────────────────────────────┐
│  Local Synaptic Gain / Plasticity / Timing                      │
│  ───────────────────────────────────────────────────────────    │
│  Fast parameters shaped by regulatory field                     │
│  • Synaptic efficacy (gain)                                     │
│  • Plasticity rate (learning rate / eta)                        │
│  • Temporal integration window                                  │
│  • Precision weights (alpha_k in REE)                           │
└───────────────────────┬─────────────────────────────────────────┘
                        ↓
┌─────────────────────────────────────────────────────────────────┐
│  E2 / E1 Predictive Dynamics                                    │
│  ───────────────────────────────────────────────────────────    │
│  Fast neuronal prediction and inference                         │
│  • E2: short-horizon prediction (z_gamma, z_beta)               │
│  • E1: long-horizon prediction (z_theta, z_delta)               │
│  • Driven by prediction errors weighted by alpha_k              │
└─────────────────────────────────────────────────────────────────┘

Key insight

Monoamines are not control knobs. They are advisory signals that bias a deeper, slower regulatory substrate. The actual control is implemented by astrocytes, which integrate these signals with local activity and metabolic constraints.

Implications for computational models

Do not treat alpha_k as an instantaneous function of monoaminergic “level.”
Model the regulatory field R(x,t) as a separate dynamical system with its own timescale.
Allow for spatial heterogeneity: different parts of L-space can have different precision/plasticity states simultaneously.
Include metabolic constraints: regulatory state should saturate or be suppressed when energy is low.

Source: docs/processed/legacy_tree/docs/astrocyte_aware_regulatory_stack/regulatory_stack_model.md

Open Questions

Q-002: What is the appropriate spatial resolution for R(x,t)?

Q-003: Should R(x,t) be a scalar or a vector (multiple regulatory dimensions)?

Q-004: How should tau_R be calibrated relative to E1/E2 timescales?

Q-005: Can offline integration be modeled as a resetting or annealing of R(x,t)?