Tantrum Bayesian Cycle
Unified Σ = variance (diagonal/PFC) + covariance (off‑diagonal/mirror neurons)
ΔS
Noodles ritual
Family ledger (posterior)
θ
Limbic flood
Raw energy (likelihood)
Σ
Compression bottleneck
Self + Shared
h(t)
Salience switch
Flip tantrum
θ′
Likelihood ↑ • Prior ↓
Soil that breathes both ways
6, 9, 12, center, 3 — if it don't fit, tog must acquit
θ — Limbic Flood (Amygdala)
Energy • Arousal • Ascending Likelihood

What Enters the Channel

Footstep tempo, sobbing pitch, cortisol and heart‑rate spikes — these are the bottom‑up likelihood gradients. They do not decide; they demand.

From θ to Σ

When arousal exceeds capacity, variance must be compressed. The PFC tries to prune and stabilize; mirror neurons invite co‑regulation.

Σ — Unified Compression
Bottleneck • Variance + Covariance • Ritual Framing

Two Modes, One Σ

Diagonal vs Off‑Diagonal (inside Σ)
• PFC (diagonal): self‑compression; glucose/oxygen spike to hold a line.
• Mirror neurons (off‑diagonal): shared compression; resonance across agents for co‑regulation.

Why Rituals Work

Rituals (e.g., noodles) reduce description length: they provide a low‑entropy policy that both agents can execute under stress, converting arousal into coordinated action.

Matrix View (Crisp)

Σ = [ var(self) (PFC) , cov(self↔other) (mirror) ; cov(other↔self) (mirror) , var(other) ]
h(t) — Salience Switch (SN)
Bifurcation • Strategy Flip • Arena

What the Flip Feels Like

The SN toggles networks: failed PFC control + high arousal → tantrum. Or, successful co‑regulation → glide path to resolution.

Outputs to ΔS

Either path writes to the ledger: meltdown then noodles vs breathing + noodles. Both become posterior evidence for next time.

ΔS — Ritual Resolution (DMN)
Posterior • Ledger • Cantabile

Posterior = What Survives

“Tantrum → noodles → calm” is encoded as a social invariant. The DMN stores what worked, not the whole storm.

Feedback to θ′

Ledgered outcomes update priors: next time, both agents expect noodles to work sooner — narrowing the channel demand at Σ.

θ′ — Bidirectional Soil
Top‑Down Priors • Bottom‑Up Likelihoods • Memory

Breathes Both Ways

Ascending signals (likelihood) climb from θ and Σ; descending structures (priors) press down from ΔS and culture. Your uncle‑scientist “ultra units” live here as compressed posteriors.

Minimal Bayesian Law

Posterior ∝ Likelihood × Prior
• Likelihood ↑: tempo, pitch, biomarkers.
• Prior ↓: “noodles soothe,” caregiving scripts, family norms.
• Posterior → ΔS: the ritual that survives.