Empowering stroke & neurological disorder recovery through real-time EMG/EEG biofeedback, LLM-driven clinical intelligence, and an integrated pharmacy management system.
- Overview
- Key Features
- System Architecture
- Tech Stack
- Getting Started
- API Reference
- User Roles & Credentials
- Future Scope
- Patent
- Contributors
- Contributing
- License
NeuroStride is an end-to-end neurorehabilitation platform built for stroke survivors, patients with cerebral palsy, traumatic brain injuries, and other neurological conditions. It bridges clinical care with cutting-edge AI and biosignal hardware to deliver personalized, data-driven recovery journeys.
The platform connects three roles — Doctor, Patient, and Pharmacist — in a unified ecosystem, backed by a real-time EMG/EEG biosignal pipeline from the Neuphony EXG Synapse board and LLM-powered intelligence via Groq.
🏆 Built for LPU Ideathon 2026 — demonstrating how AI can transform neurological rehabilitation in India.
- View and manage all assigned patients
- Create personalized exercise plans with repetitions, sets, and goals
- Write digital prescriptions that auto-sync to the pharmacy system
- Review and approve AI-generated progress reports with one click
- Download clinical reports as formatted Word documents
- Live Sensor Feed — Real-time EMG waveform visualization with intent detection
- Guided Exercise Sessions — Start/end sessions with automatic metric logging
- AI Chatbot — Multilingual rehab assistant (English, Hindi) powered by Groq LLM
- Progress Dashboard — Recharts-based analytics of rehab trends and milestones
- Prescription History — Track medications and pharmacy order status
- Auto-receives orders from doctor prescriptions
- Inventory management with low-stock alerts
- Process and dispense orders with status tracking
- Generate itemized tax invoices (PDF/DOCX) with GST calculation
- Neuphony EXG Synapse (ESP32) — Reads raw EMG/EEG/ECG signals via USB serial
- Python-side IIR Filters — Mirror of Synapse.h filter coefficients for offline processing
- FFT Band Analyser — Delta / Theta / Alpha / Beta / Gamma power in real-time
- Servo Arm Control — Intent-triggered robotic hand (for assistive therapy)
- WebSocket Bridge — Streams sensor data to the frontend at 20 fps
- Simulation Mode — Realistic synthetic EMG signal for demos without hardware
┌─────────────────────────────────────────────────────────────────┐
│ NeuroStride │
│ │
│ ┌──────────────┐ ┌─────────────────┐ ┌──────────────┐ │
│ │ Next.js 14 │────▶│ FastAPI Core │────│ SQLite / │ │
│ │ Frontend │◀────│ (Port 8000) │ │ PostgreSQL │ │
│ │ (Port 3000) │ └────────┬────────┘ └──────────────┘ │
│ └──────────────┘ │ │
│ ┌─────┴──────┐ │
│ ┌────┴───┐ ┌────┴────┐ │
│ │ Groq │ │Langfuse │ │
│ │ LLM │ │Observ. │ │
│ └────────┘ └─────────┘ │
│ │
│ ┌─────────────────────────────────────────────────────────┐ │
│ │ Hardware Layer (Optional) │ │
│ │ Neuphony EXG ──serial──▶ neuphony_bridge.py │ │
│ │ (ESP32) WebSocket :8765 ──▶ Frontend │ │
│ │ Servo Controller ──▶ Arm │ │
│ └─────────────────────────────────────────────────────────┘ │
└─────────────────────────────────────────────────────────────────┘
| Layer | Technology |
|---|---|
| Backend | Python 3.11, FastAPI, SQLAlchemy, SQLite/PostgreSQL |
| Auth | JWT (python-jose), Passlib/bcrypt |
| AI | Groq (Llama 3.3 / Mixtral), Langfuse observability |
| Frontend | Next.js 14, React 18, Recharts, Axios |
| Hardware | Neuphony EXG Synapse (ESP32), PySerial, WebSockets |
| Docs | python-docx (DOCX report & invoice generation) |
| Deploy | Render (backend), Vercel (frontend) |
| Tool | Version | Download |
|---|---|---|
| Python | 3.11+ | python.org |
| Node.js | 20+ | nodejs.org |
| Git | any | git-scm.com |
# 1. Clone the repo
git clone https://github.com/LTPratham/neurostride.git
cd neurostride
# 2. Create a virtual environment
cd backend
python -m venv venv
.\venv\Scripts\activate # Windows
# source venv/bin/activate # Linux / macOS
# 3. Install dependencies
pip install -r requirements.txt
# 4. Configure environment variables
copy .env.example .env # Windows
# cp .env.example .env # Linux / macOS
# → Edit .env and fill in your GROQ_API_KEY, SECRET_KEY, etc.
# 5. Seed demo data (first run only)
python seed_data.py
# 6. Start the API server
uvicorn main:app --reload --port 8000API is running at: http://localhost:8000
Swagger docs: http://localhost:8000/docs
# In a new terminal
cd neurostride/frontend
npm install
npm run devFrontend at: http://localhost:3000
Skip this if you want to run in simulation mode — the platform works fully without physical hardware.
- Download Arduino IDE
- Clone the Neuphony EXG Synapse repo
- Open
EXG-Synapse/src/EMG/serial/EMG_serial.ino - Select Board:
ESP32 Dev Module - Select Port: your COM port (check Device Manager → Ports)
- Upload the sketch
- Verify in Serial Monitor at 115200 baud — you should see numbers streaming
cd neurostride/hardware
# With real hardware:
python neuphony_bridge.py --port COM3 --servo COM4
# Without hardware (simulation mode):
python neuphony_bridge.py --simulate
# Auto-detect port:
python neuphony_bridge.pySensor WebSocket: ws://localhost:8765
| Endpoint | Method | Description |
|---|---|---|
/api/auth/register |
POST | Register a new user |
/api/auth/login |
POST | Login and receive JWT |
/api/auth/me |
GET | Get current user info |
/api/patients |
GET | List all patients (Doctor/Pharmacist) |
/api/patients/{id} |
GET/PUT | Get or update patient profile |
/api/sessions |
POST | Start a rehab session |
/api/sessions/{id}/end |
PUT | End session with metrics |
/api/exercise-plans |
POST | Create exercise plan (Doctor) |
/api/prescriptions |
POST | Create prescription (Doctor) |
/api/pharmacy/orders |
GET | List pharmacy orders |
/api/pharmacy/inventory |
GET | Get medicine inventory |
/api/reports/patient/{id} |
GET | Get progress reports |
/api/reports/{id}/download |
GET | Download report as DOCX |
/ws/sensor/{patient_id} |
WS | Live sensor data stream |
Full interactive documentation: http://localhost:8000/docs
After running python seed_data.py:
| Role | Password | |
|---|---|---|
| Doctor | dr.sharma@neurostride.in |
doctor123 |
| Pharmacist | pharmacy@neurostride.in |
pharmacy123 |
| Patient 1 | ravi@neurostride.in |
patient123 |
| Patient 2 | sunita@neurostride.in |
patient123 |
| Patient 3 | arjun@neurostride.in |
patient123 |
NeuroStride is built with extensibility at its core. Here's the roadmap for what comes next:
- Adaptive Exercise Plans — LLM automatically adjusts difficulty based on session performance trends
- Predictive Recovery Timelines — ML model trained on session data to forecast recovery milestones
- Anomaly Detection — Alert doctors when EMG/EEG patterns deviate significantly from baseline
- Voice-Controlled Interface — Hands-free navigation for patients with limited motor function
- Multi-modal AI — Integrate computer vision (MediaPipe pose estimation) alongside EMG for form scoring
- React Native App — Cross-platform mobile app for patients to log sessions remotely
- Offline Mode — PWA support so patients can use the app without internet
- Multilingual Expansion — Add Tamil, Telugu, Bengali, Marathi support to the AI assistant
- Accessibility Audit — WCAG 2.1 AA compliance for visually impaired users
- Multi-Channel EEG Support — Upgrade to 8-channel EXG board for full brain mapping
- Wireless Streaming — ESP32 Wi-Fi/BLE mode to eliminate USB cable dependency
- 3D-Printed Exoskeleton — Full-hand exoskeleton with individual finger servo control
- Wearable Form Factor — Miniaturized PCB that attaches to the patient's forearm
- ECG Monitoring — Heart rate variability (HRV) tracking during rehab sessions
- ABDM Integration — Link with India's Ayushman Bharat Digital Mission for health ID interoperability
- HIPAA / DPDPA Compliance — Encrypt PHI at rest, audit logs, data retention policies
- Telemedicine Module — In-platform video consultation between doctor and patient
- HL7 FHIR Export — Standardized health data export for interoperability with hospital systems
- Clinical Trial Dashboard — Aggregate de-identified data for research and outcome studies
- Kubernetes Deployment — Helm chart for production-grade scaling on GKE/EKS
- CI/CD Pipeline — GitHub Actions for automated testing, linting, and deployment
- Real-time Notifications — WebPush notifications for prescription updates and session reminders
- Multi-tenant SaaS — Support multiple hospitals with isolated data namespaces
- Analytics Dashboard — Aggregated, anonymized population-level rehab outcomes
NeuroStride's core technology — including its real-time EMG/EEG biofeedback pipeline and AI-assisted neurorehabilitation methodology — is protected by a published patent.
| Field | Details |
|---|---|
| Inventor | Prathmesh (LTPratham) |
| Status | Published |
| Document | docs/NeuroStride_Patent_Published.pdf |
The full patent document is available in the docs/ folder of this repository.
Contributions are welcome! Please read CONTRIBUTING.md for guidelines on branching, commit conventions, and the PR process.
Please review our SECURITY.md before reporting vulnerabilities. Do not open public issues for security bugs.
 Prathmesh Founder · Full-Stack · AI · Hardware |
 Bhavya Verma Co-Developer · Contributor |
See CONTRIBUTORS.md for the full list.
This project is licensed under the MIT License.
Built with ❤️ at LPU Ideathon 2026
"Restoring movement. Restoring life."