Skip to content

aoxilus/MindWaveRF

Repository files navigation

🧠 MindWaveRF

Turn a vintage NeuroSky MindWave RF headset into an open BCI lab in the browser.
Live EEG · blink games · Attention / Meditation studies · JSON for AI
by aoxilus 🥑

🇪🇸 Español · Repo · Contribute

topic bci platform stack license


💡 What is this?

MindWaveRF is an open toolkit so students, makers, and researchers can actually use the white NeuroSky MindWave (RF + USB dongle) — the old wireless headset that often dies behind Code 52 drivers, dead downloads, and ThinkGear Connector drama.

We bridge the dongle → WebSocket → browser apps: see your brainwaves, train Attention/Meditation, fire actions with blinks, and export sessions for math / AI analysis.

Not magic mind-reading. Honest 1-channel BCI: blink = best voluntary command, Attention = slow continuous control, raw EEG for science & viz.


🎧 What hardware?

Piece Notes
🧠 NeuroSky MindWave (white, AAA) Forehead sensor + ear clip
📡 MindWave Wireless USB Adapter RF dongle (VID_1A86 / PID_7523, often shows as CH340)
💻 Windows 10/11 PC Signed CH340 / MindWave driver (we ship Drivers/)

Also works with the ThinkGear protocol ideas used across the neurosky-mindwave community — but we optimize for RF dongle, not only Mobile Bluetooth.


🎯 Who does it help?

You are… You get…
🎓 Student / lab Cheap BCI demos without $2k caps
🕹️ Game / UX maker Blink shoot · Attention power bar patterns
🔬 Researcher Session JSON + FFT + band logs for analysis
🛠️ Driver survivor Docs for Secure Boot, COM, baud 115200
🤖 AI tinkerer Exports ready to paste into another model

What we’re not claiming: decoding color, “thought left/right”, or secret thoughts from one forehead electrode. We keep stimulus labels so you can test correlations — and we publish null results when they’re null.


🔍 What we’re looking for (mission)

  1. Reliable RF link on modern Windows (pair fast, stay blue).
  2. Honest metrics — Attention / Meditation / blink / raw / FFT you can trust on-screen.
  3. Study protocols that teach what 1-channel BCI can and cannot do.
  4. Open algorithms (Kalman, blink slope detectors, short Attention forecast) that improve the live graph without snake oil.
  5. Contributors who document hardware quirks so the next person doesn’t lose a weekend.

Community lessons we follow: docs/studies/github-neurosky-landscape.md · AI ideas: docs/studies/ai-algorithms-for-wave.md


✨ What’s inside

🌐 App What it does
wave.html 📈 Live EEG · Attention / Meditation · raw (avg×5) · FFT · bands · blink · concentrado / relajado
attention.html 🎯 Personal Attention threshold
calibrate.html 👁️ Reading + blink · shape · side · color study (JSON for AI)
portal.html ⚽ Mental goal — Attention = power, blink = shoot
index.html 🕹️ Monkey Run mind game
Drivers/ 🔌 USB dongle driver (CH340)
server/ 🌉 Serial / TGC → WebSocket :13855

🚀 Quick start

📘 Full guide: docs/connect.md

npm install
npm run serial -- COM18 115200   # 🔌 change COM if needed
npm run waves                    # 📈 http://localhost:5173/wave.html

✅ Checklist

  1. 🔌 Plug the RF USB dongle (COM in Device Manager)
  2. 🎧 Headset ON → blue LED, forehead sensor + ear clip
  3. 🌉 Run the bridge + open wave.html

🔴 Red LED = powered but not linked — move closer.
🔋 Won’t power on? → docs/headset-no-power.md

⚠️ With signed USB-SERIAL CH340, stable ThinkGear was 115200 (not 57600) on our test PC.


🤝 Contribute — we want your help

Stars, issues, and PRs are welcome. Especially:

🙋 Help with… Examples
📡 RF pairing reliability Faster 0xC2 strategies, COM race fixes
📈 Signal algorithms Kalman / forecast line, blink slope detector, α-power relax
🧪 Study designs Catch trials, longer reading windows, CSV exports
🪟 Windows drivers Secure Boot / Code 52 war stories & INF tips
🌍 Docs / i18n Clearer guides, more languages
🕹️ Games Attention → Monkey Run speed (community runner pattern)

How to start

  1. Fork → branch → PR against main
  2. Keep changes small; link an issue if you can
  3. Don’t commit huge study/*.json waveLogs (gitignored) — commit analysis under docs/studies/
  4. Be honest in docs: say what failed

Open tasks: docs/tasks.md · Function map: docs/function-vectors.md

Good first issues ideas:
- Attention → speed in Monkey Run
- Dashed Kalman forecast on wave.html
- Double-blink as second command
- Repo description / screenshots / GIF of blue LED → streaming

🧪 BCI studies

Command Opens
npm run attention 🎯 Attention calibration
npm run calibrate 👁️ Blink / reading / figure study
npm run portal ⚽ Mental goal
npm run waves 📈 EEG + FFT monitor

📤 Each study has Export JSON (AI) — paste into another AI for math analysis.
More: docs/studies/README.md


🧰 Scripts

Command Action
npm run serial -- COMx 115200 🌉 Dongle → ws://127.0.0.1:13855
npm run waves 📈 Open wave.html
npm run attention 🎯 Attention cal
npm run calibrate 👁️ Blink study
npm run portal ⚽ Portería
npm run dev 🕹️ Vite (all apps)
npm run bridge 🔗 Official TGC → WebSocket

⛔ Don’t run ThinkGear Connector and npm run serial on the same COM at once.


📚 Docs

Doc Topic
docs/connect.md 🔌 How to connect
docs/function-vectors.md 🧬 Condensed function map
docs/tasks.md ✅ Closed / open tasks
docs/studies/README.md 🧪 Studies + AI export
docs/studies/github-neurosky-landscape.md 🌍 GitHub landscape
docs/studies/ai-algorithms-for-wave.md 🤖 Algorithms to try
docs/findings.md 📝 Local findings
docs/thinkgear-connector.md 🔗 TGC
docs/windows-11-secure-boot.md 🛡️ Secure Boot / Code 52
docs/rf-pairing-and-testing.md 📡 RF pairing

📀 Driver

Name MindWave USB Adapter
Version 3.11.2015.08 (2015-08-03)
USB ID VID_1A86 / PID_7523 (CH340)
Files Drivers/

Source: sieuwe1/ArduMind Drivers. Official NeuroSky downloads are often offline.


⚖️ License / disclaimer

Driver binaries and NeuroSky protocol as-is for personal/educational use. No warranty.
You are responsible for Windows driver-signing policy on your PC.

Hardware/firmware © NeuroSky / WCH.


Made with 🥑 by aoxilus
If this saved you a weekend of COM ports — star the repo and open an issue with your headset + Windows build.

About

Open BCI toolkit for NeuroSky MindWave RF: live EEG, blink games, Attention studies, JSON for AI. by aoxilus

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages