France Bans Social Media for Kids Under 15 — And Why We Built a VR Vision Trainer Instead
France just became the first European country to ban social media for under‑15s. The law targets algorithmic feeds that keep children scrolling for hours. The public discourse quickly collapses into “screens bad.”
But as a dev team building in VR health, we see a missing distinction: passive consumption ≠ active therapeutic training. A child staring at an endless feed is not the same child who is deliberately exercising a weak visual pathway under controlled, measurable conditions.
The Problem: Amblyopia & The Patch
Amblyopia (lazy eye) affects roughly 3 % of children worldwide. Standard care still relies on occlusion therapy — patching the strong eye for several hours a day. Compliance drops sharply after age seven; kids resist the patch, parents burn out, and the critical window for cortical plasticity narrows. Even when families persist, the patch offers no feedback on whether the amblyopic eye is actually improving.
The Science: Dichoptic Training
Dichoptic presentation shows different contrast or luminance to each eye. When the brain receives a balanced binocular signal, the suppressed eye is forced to contribute, re‑engaging cortical pathways that were previously ignored. Research demonstrates that this approach can work at any age, provided the per‑eye rendering is precise and the stimulus drives primary visual cortex (V1) neurons — oriented Gabor patterns are a proven V1 driver.
The Build: AmblyoPunch (Meta Quest)
We translated that science into a four‑stage dichoptic protocol wrapped in a game loop that feels like play, not a clinic visit.
1. Calibration & Config
A staircase procedure measures each eye’s contrast threshold. Wrist‑ and wall‑mounted menus let the clinician or caregiver set:
- Lazy eye (L/R) — a wrong choice flips the clinical effect.
- Severity (Mild / Moderate / Severe) — scales target size, speed, and stage duration.
- Strabismus type (Esotropia / Exotropia) — determines the vergence offset used in Stage 4.
All settings are stored locally; no cloud accounts, no telemetry.
2. Stage 1 — Monocular Warm‑up
Gabor‑coined coins and red spikes render only to the amblyopic eye. The dominant eye sees a blank field, eliminating inter‑ocular competition and letting the weak eye build cortical tracking confidence.
3. Stage 2 — Suppression Breaking
Both eyes receive the scene, but the dominant eye’s alpha and contrast are heavily crushed. The amblyopic eye carries the usable signal, forcing the brain to rely on it.
4. Stage 3 — Rebalancing
Across a series of sub‑steps the dominant eye’s brightness and contrast are gradually restored. The amblyopic eye remains online, learning to share the visual workload.
5. Stage 4 — Fusion Training
Micro‑disparities placed inside Panum’s fusional area require true binocular fusion to hit targets. Esotropia and exotropia profiles receive tailored vergence offsets so the demanded convergence/divergence matches the child’s specific deviation.
Gameplay Loop
- Core mechanic: punch oriented Gabor coins (V1 drivers) and dodge red spikes.
- Lives: three hearts; loss ends the session.
- Input: hand‑tracking plus optional controllers for accessibility.
- Auto‑difficulty: velocity ↑, coin size ↓, spike size ↑ as performance improves.
- World progression: village → city → space → moon → Mars → Venus, giving a sense of journey.
Progression Logic
A rolling window of the last 10 outcomes (punch or dodge) drives advancement:
- ≥ 70 % success → advance to next sub‑step or world.
- < 40 % → risk demotion, protecting against frustration.
- Strabismus profiles require ≥ 80 % before unlocking Stage 4, ensuring stable fusion before added vergence demand.
Tech Stack Notes
- Engine: Unity 2022 LTS with Meta Quest SDK.
- Per‑eye rendering: OVRCameraRig dual‑pass pipeline.
- Gabor shaders: procedurally generated at runtime, guaranteeing exact spatial frequency and orientation.
- Data model: pure local JSON; no network calls, no analytics, GDPR‑by‑design.
Safety Positioning
AmblyoPunch is marketed as a training/assistive game, not a medical device. It is intended to complement, not replace, professional orthoptic or ophthalmologic care. Clinicians can use the built‑in calibration data to monitor adherence and adjust patching schedules.
France banned the algorithm that hijacks attention. We built the alternative that directs attention toward measurable visual improvement.
👉 https://www.meta.com/en-gb/experiences/amblyopunch/1239507485902689/
Top comments (0)