Project-Synapse-Real-Time-Brainwave-Decoder-with-Android

mikey

Copyright © 2025 Yogesh R. Chauhan

Project Synapse – Brainwave Decoder

This project is licensed for personal and educational use only.

Commercial use, resale, or modification for profit is strictly prohibited.

Unauthorized use will result in legal action and takedown notices.

Project Synapse: Real-Time Brainwave Decoder with Android

Build a DIY EEG system to capture and decode your brain’s electrical activity into meaningful words in real-time — all using own eeg circuit,Arduino, Android, and Arch Linux inside Termux.

📸 Electrode Placement

Use 3 electrodes based on the following 10-20 system positions:

Pz → T1 (center forehead)
Fpz → T3 (above nose, optional)
Mastoid → T2 (behind ear, ground/reference)

Electrode Placement

place electrodes according to below image

Pz → T1 (center forehead)
Fpz → T3 (behind the ear at right side of face)
Mastoid → T2 (as image front&below of ear)

Electrode Placement

i tried with this setup of electrodes and it’s working perfectly 👍

🧰 Required Hardware

Arduino Uno board
EEG Circuit (NE5534AP/TL081 based)
Electrodes (snap gel or dry)
Android phone (USB OTG supported)
USB OTG cable

🪲 Circuit Diagram

The following hand-drawn schematic shows the EEG circuit + power supply:

EEG Circuit

🛠️ Step-by-Step Setup Guide

🔌 1. Build the EEG Hardware

Follow the circuit in the image above and connect:

T1 → Pz (EEG active)
T2 → Mastoid (Reference)
T3 → Fpz (optional, can be left unconnected)

📱 2. Install Termux & Packages on Android

📲 Download Termux (F-Droid)

termux-setup-storage
pkg update && pkg upgrade -y
pkg install wget git curl termux-api -y
pkg install tur-repo x11-repo -y
pkg install termux-x11-nightly pulseaudio -y

🐧 3. Install Arch Linux via proot-distro

pkg install proot-distro -y
proot-distro install archlinux
proot-distro login archlinux

Inside Arch:

pacman -Syu python python-pip git base-devel wget --noconfirm

🧪 4. Setup Python Virtual Environment with Pyenv and install tensorflow

Follow this repository for installing pyenv and tensorflow in Termux or Arch or use below commands.

Tensorflow in Android

Create and activate the virtual environment:

pyenv install 3.10.13
pyenv virtualenv 3.10.13 tf-env
pyenv activate tf-env

📦 5. Install Python Dependencies

pip install --upgrade pip
pip install numpy==1.24.3 pandas scipy joblib scikit-learn tensorflow==2.13.1

🔌 6. Upload Arduino Sketch to Uno

Use the following code:

// EEG P300 Arduino v1.0 — sends raw samples at FS via Serial
#define EEG_PIN A0
#define FS 250
#define BUFFER_SIZE 128
volatile uint8_t buf[BUFFER_SIZE];
volatile int head = 0, tail = 0;
void setup() {
  Serial.begin(9600);
  noInterrupts();
  TCCR1A = 0; TCCR1B = 0; TCNT1 = 0;
  OCR1A = (16000000/(8*FS)) - 1;
  TCCR1B |= (1 << WGM12) | (1 << CS11);
  TIMSK1 |= (1 << OCIE1A);
  interrupts();
}
ISR(TIMER1_COMPA_vect) {
  int val = analogRead(EEG_PIN);
  uint8_t sample = val >> 2;
  buf[head] = sample | 0x80;
  head = (head + 1) % BUFFER_SIZE;
}
void loop() {
  while (tail != head) {
    Serial.write(buf[tail]);
    tail = (tail + 1) % BUFFER_SIZE;
  }
}

📲 7. Connect Android with Arduino (via OTG)

Install TCPUART APP from tcpuart

Set baud rate: 9600

Press Connect

Allow USB permission & select Server Mode

set Port: 8080

Press Start

To check Arduino’s signal install:

wget https://raw.githubusercontent.com/mikey-7x/Project-Synapse-Real-Time-Brainwave-Decoder-with-Android/refs/heads/main/q3.py

Run:

python q3.py

if any numerical values shown that meas signal coming perfectly.now you can run actual abd57.py (advance brainwave decoder 57) script.👍

☣️ 8. Run EEG Decoder Software in archlinux in termux

Clone this repo and activate virtualenv:

cd ~
git clone https://github.com/mikey-7x/Project-Synapse-Real-Time-Brainwave-Decoder-with-Android.git
cd Project-Synapse-Real-Time-Brainwave-Decoder-with-Android
pyenv activate tf-env

Run the script:

python abd57.py

You will see:

🧬 Quantum-Level EEG Decoder

[1] Record New Alphabet Samples

[2] Train from Existing Directory

[3] Predict in Real-Time

[4] Switch Model (Now: LSTM)

[5] Clear History

[6] Record Custom Word

[7] Discover Unique Patterns

[8] Exit

🎓 How to Use

Use [6] Record Custom Word → enter a word like yes or go

Think of it strongly during the recording session

Once training completes automatically, go to [3] Predict in Real-Time

⚠️ Note: Accuracy is best with 10–15 trained words only. More than that causes model confusion. Strong mental focus during training/predictions improves accuracy even above 20 trained words 🪽

📂 Files Included

File:Description

abd57.py: EEG decoder, trainer & real-time predictor

abd.ino/abd.hex: Arduino EEG sampling code

README.md: This documentation

🛑 Disclaimer

This is not a medical device. It’s an experimental research tool for signal processing and brain-computer interface (BCI) exploration.

📜 License

This repository is released under a custom license for educational and academic use only.

🚫 Commercial use or resale of any part (code, circuit, models, diagrams) is strictly prohibited.
✅ Use in college/university research, projects, and learning is permitted.
📩 Contact the author for special permissions or collaboration.

See the full LICENSE file for details.

☕Auther

Made with ❤️ by mikey-7x For research & educational use only.🚀🔥

other repository

This site is open source. Improve this page.