EpilepsyPredictor is an AI-based system designed to predict and detect epileptic seizures in real-time using EMG data and Machine learning.
Developed by NeuroVectorLab — advancing accessible neurotechnology.
EpilepsyPredictor leverages convolutional neural networks and signal processing to analyze EEG & EMG activity and identify preictal patterns.
The goal is to provide early alerts for patients and clinicians, improving safety and preventive treatment planning.
- ⚡ Real-time EMG signal processing
- 🧩 Machine learning inference (Random forest)
- 📊 Visualization dashboard for brain activity and alerts
- 🧱 Modular FastAPI backend with Streamlit frontend
- ☁️ Deployment-ready for Railway, Hugging Face Spaces, or local environments
├── software/ # Python & C
├── hardware/ # Device
├── models/ # Pretrained AI weights
├── data/ # Signal datasets
├── docs/ # Documentation and reports
├── assets/ # Logo, banners, visual materials
├── LICENSE (MIT personal no-comercial).txt / License
├── LICENSE-COMMERCIAL.txt / License commercial
└── README.md # You are here
# Clone repository
git clone https://github.com/NeuroVectorLab/EpilepsyPredictor.git
cd EpilepsyPredictorArchitecture: Random forest (Best for this Hardware)
Input: EMG time-series data (1 channel)
Output: Seizure probability (Normal, Preictal, Ictal)
Dataset: CHB-MIT Scalp EEG Database
Frameworks: PyTorch, NumPy, SciPy, Scikit-learn
Complete project documentation is available in the /docs folder:
hardware_description.md → Hardware structure
software_description.md → Hardware structure
user_guide.md → User Guide
[x] Base model training.
[x] Inference on Device.
[x] Hardware ready with model.
[ ] Clinical validation phase.
If you believe in open neurotechnology, you can help us grow:
💌 Contact: neurovectorlab@gmail.com
Name Role Contact
Jorge Founder, AI Research & Development GitHub
NeuroVectorLab Team Design & Neuroscience Collaboration neurovectorlab@gmail.com
Special thanks to the following organizations for their valuable support and open collaboration in advancing neurotechnology research:
- DFRobot — for providing high-quality hardware components used in prototype testing and data acquisition.
- Upside Down Labs — for their innovative biosensing hardware, which greatly contributed to EEG and EMG experimentation.
- Massachusetts Institute of Technology (MIT) — for publishing open-access documentation and datasets that have been essential to the scientific foundation of this research.
- A. Djemal, M. Alabed, A. Alkassih, Wearable Electromyography Classification of Epileptic Seizures, Frontiers in Neuroscience, 2023. ** Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC10326816/ [Accessed 5 Nov. 2025].
- S. K. Patel, R. B. Joshi, Development of EMG Based Epileptic Seizure Detection Using LSTM Approach, ResearchGate Preprint, 2025. Available at: https://www.researchgate.net/publication/392986299_Development_of_EMG_based_Epileptic_Seizure_Detection_using_LSTM_Approach [Accessed 5 Nov. 2025].
- R. Babu, P. K. Sharma, S. R. Nair, EEG and EMG Signal Integration Using Machine Learning for Detection of Epileptic Seizures, in: Advances in Artificial Intelligence and Data Engineering, Springer, Singapore, 2025, pp. 75–86. Available at: https://link.springer.com/chapter/10.1007/978-981-96-9967-4_7 [Accessed 5 Nov. 2025].
Your contributions make accessible neuroscience possible. Thank you for inspiring the next generation of open neurotechnology.
This project uses a dual license:
MIT License — for personal and research use
Enterprise License (custom) — for commercial applications (contact us for details)
NeuroVectorLab provides this software “as is”, without warranty of any kind.
🌍 Website: neurovectorlab.github.io
🧠 Organization: NeuroVectorLab on GitHub
💼 Linkedin
“Predicting seizures. Protecting lives.”
© 2025 NeuroVectorLab