Calm Space — The Regulation Doctrine
*Why Calm Space exists, what it must never do, and the engine that decides what the child meets and when. This document governs every build decision. Where it conflicts with novelty, engagement, or "fun," this document wins.*
*Grounded in the current evidence base: dopaminergic hypo-responsivity in autism (reduced dopamine release in the mesocorticolimbic pathway; diminished nucleus accumbens responsiveness), the psychomotor-regulation model of autism (difficulty initiating, inhibiting, and sustaining movement, thought, and emotion), and the evidence that sensory-integration and self-regulation interventions in naturalistic home settings meet the bar for evidence-based practice. Not medical advice; for review by an OT and a developmental clinician before scaling.*
---
1. Why we are doing this — the one paragraph that governs all others
This child's brain is not lazy, not stubborn, and not broken. It is **confused and dysregulated**: it struggles to interpret and sequence sound, light, and movement, its reward system responds weakly and unpredictably, and it cannot reliably tell anyone how it feels. Every other thing we build serves one prior goal — **bring the confused brain into balance**. A regulated brain can attend. An attending brain can connect. A connecting brain discovers purpose. We do not start with learning, motivation, or speech, because none of those are reachable from dysregulation. We start by making the world calm enough to process, predictable enough to trust, and responsive enough to feel worth acting on. Regulation is the foundation; everything else is built on it, in order, or not at all.
---
2. The line that separates us from entertainment
> **We regulate; we do not stimulate. Every good feeling here is earned through competence, never delivered through spectacle.**
A confused, hypo-responsive reward system does not need flooding — flooding is noise to a brain already drowning in noise. It needs **small, reliable, predictable signals it can actually process**. The dopamine we want is the quiet kind that comes from *"I made that happen"* — competence, mastery, agency. The dopamine we refuse is the loud kind from surprise, escalation, and sensory spectacle — the "rush" that entertainment apps engineer and that, for this child, is not pleasure but a step toward overload and the behavioural collapse that follows.
**The test for any activity, sound, animation, or reward:** Does this regulate, or does it arouse? If it spikes, surprises, escalates, or exists to be exciting — it does not ship.
---
3. The five hard rules (non-negotiable)
1. **No variance.** Same input → same output, identically, every single time. No randomised tones, colours, timings, or rewards. Predictability is not a preference here; it is the mechanism. A brain that cannot process variance must never be asked to. *(This corrects the prototype's random Ripples pitch and any "delightful variation" instinct.)*
2. **No arousal spikes.** No bright "success" chimes, no upward musical leaps, no celebratory bursts, no escalating feedback. Acknowledgment is warm, low, brief, and the same as last time. *(This corrects the prototype's Rhyme reward jumping to a bright high tone.)*
3. **No unregulated reward.** Reward is tied only to the child's own action and is proportionate, gentle, and immediate. We never reward to *pull* the child into more activity. We never build a loop whose purpose is continued engagement.
4. **Sensory-overload prevention is priority one — above engagement, above progress, above everything.** The system's first duty in every moment is to avoid tipping the child toward overload. When in doubt, it does less, dims more, slows down, goes quiet. A boring, safe moment is always preferable to an exciting, risky one.
5. **Regulate before anything.** Nothing is offered to a dysregulated child except the path back to calm. Learning, motivation, communication, and play are only ever offered from a regulated baseline — never as a way to *achieve* one.
---
4. Reading the child who cannot tell us — the safety-first sensing model
The child is non-verbal and cannot describe their state. We must therefore **infer arousal from behaviour, and when inference is uncertain, default to the safest action: down-regulation toward calm.** We never assume a direction. We read, then act.
**Signals the system can honestly observe (no camera required for v1):**
- **Touch tempo and pressure** — rapid, hard, scattered touches suggest rising arousal (over-arousal). Absence of touch over time may suggest withdrawal (under-arousal) — but may equally mean contented watching, which is success. The system must not confuse stillness with distress.
- **Erratic navigation** — fast jumping between activities suggests dysregulation; settled dwelling suggests regulation.
- **Repetition pattern** — returning to the same calm thing is self-regulation working; we honour it, never interrupt it.
**The two directions, and the safe response to each:**
| Reading | What it may mean | The safe response |
|---|---|---|
| **Rising / over-aroused** — fast, hard, scattered touch; rapid jumping | Overwhelm building toward meltdown | Down-regulate: fade to the breathing calm, dim, slow, quiet. (This is the existing auto-settle — keep and strengthen.) |
| **Flat / under-aroused** — long withdrawal *with* signs of disengagement | Shutting down | Gentle, low-demand *invitation* only — never a push. A single soft, slow, optional movement of colour that asks nothing. If ignored, it stops and returns to calm. We never escalate to rouse. |
| **Uncertain** — we cannot tell | unknown | **Default to calm.** Always. The cost of wrongly calming a content child is a quiet moment. The cost of wrongly rousing an overwhelmed one is a meltdown. We never take that risk. |
**The cardinal sensing rule:** *When the system is unsure whether the child needs calming down or bringing up, it calms. Down-regulation is always the safe default, because over-arousal causes the behavioural harm we most need to prevent, and a calm child is never harmed by more calm.*
A camera-based affect read (planned later) only ever *adds confidence to down-regulation decisions* — it is never used to justify rousing or stimulating a child. And it passes through the Mother's Review Gate like everything else.
---
5. The sequencing engine — what the child meets, and when
Today the prototype offers six doors at once. A dysregulated brain should not choose from six. The engine's job is to **meet the child at their current state and offer only the next regulated step** — the "confused brain, bring balance" principle made into logic.
**State the child is in → what the engine offers:**
- **Unknown / just arrived** → Watch only. Pure calm. Ask nothing. (Right to Passivity; safety baseline.)
- **Regulated and settled** → the next rung is *gently* available — but never auto-launched. The child reaches; the door is open; we do not pull them through. (Ladder rungs from the mother's letter: Watch → Ripples → preference → shared attention → Talk → Rhyme.)
- **Rising arousal** → all forward doors quietly recede; only the path to calm remains. Down-regulate.
- **Withdrawn** → calm remains; one low-demand optional invitation may appear; if unanswered, it dissolves.
**Engine principles:**
- The engine **reduces choice under stress and widens it only under calm.** Choice is a privilege of regulation, not a constant.
- The engine **never sequences toward "more engagement."** It sequences toward *the right amount of regulated input for this moment* — which is sometimes less, sometimes nothing.
- Forward movement (up the ladder) is **offered, never imposed, and only ever from a regulated baseline.** A child who stays on Watch for months is succeeding.
- Every forward step is also a **purpose** step: each rung exists because it builds a real capacity (agency, shared attention, communication, voice) — never because it is novel or fun. Motivation here means *"this action matters and I can do it,"* not *"this is exciting."*
---
6. Motivation and purpose — the honest version
You named the real goal: the child should *feel motivated, with a purpose, in a regulated way.* Here is what that means mechanically, and what it must never become:
- **Motivation comes from competence, not arousal.** "I touched, and it answered — the same way, again" is the engine of motivation for this child. Mastery of a predictable thing is motivating precisely *because* it is predictable.
- **Purpose comes from agency that works.** "I told you something and you understood" (Talk), "I made a sound and it was met with warmth" (Rhyme) — these are purpose. They build because each one teaches *my action changes my world.*
- **What it must never become:** a reward schedule that pulls the child into more screen time; a variable-reward loop; a difficulty curve that chases engagement; anything that treats the child's attention as a metric to maximise. We are not optimising engagement. We are building a regulated brain that discovers it has agency.
---
7. What this changes in the build — immediate
1. **Ripples:** one fixed tone, identical every time. Remove all randomisation. *(Hard rule 1.)*
2. **Rhyme:** replace the bright high "success" chime with a warm, low, calm acknowledgment — same every time. *(Hard rule 2.)*
3. **Auto-settle:** keep and strengthen as the core down-regulation circuit; make it the universal safe default the engine falls back to whenever a reading is uncertain. *(Section 4.)*
4. **Sequencing engine:** specify and build the arousal-aware engine (Section 5) so the child is met at their state, not handed six doors. This is the next real build, and the highest-leverage one.
5. **Audit every sound, animation, and transition** against the Section 2 test — regulate or arouse — and remove or soften anything that arouses.
---
8. The check (this document's version of the constitution's check)
Before anything ships, one question governs:
> **Does this help a confused brain find balance — or does it ask a confused brain to process more than it can?**
If it regulates, calms, makes the world more predictable, or builds agency gently — it ships. If it stimulates, surprises, escalates, or exists to engage — it does not, no matter how good it looks. *Sensory safety and regulation always outrank engagement, progress, and delight. Always. That is the whole point.*