DysphaGrip — Making Swallowing Monitoring More Accessible
Inspiration
For many stroke survivors, something as simple as swallowing water can become dangerous. Dysphagia, or difficulty swallowing, affects approximately 42% of acute stroke patients and significantly increases the risk of pneumonia and mortality.
Despite these risks, swallowing is difficult to monitor continuously outside specialized clinical environments. Standard diagnostic procedures often require expensive imaging equipment, trained specialists, and hospital visits.
We wanted to explore a simple question: What if affordable, wearable sensors could make swallowing-related signals easier to capture, visualize, and study?
That question inspired DysphaGrip.
What It Does
DysphaGrip is a low-cost, wearable research prototype designed to capture and visualize the physical signals associated with swallowing.
Our system combines two sensing technologies:
- Contact microphone: Captures vibrations from the surface of the throat during swallowing.
- MPU-6050 accelerometer and gyroscope: Measures neck movement across six axes.
By synchronizing these signals, DysphaGrip provides a more comprehensive picture of swallowing activity than either sensor alone.
Our interactive dashboard displays real-time waveforms, spectrograms, and motion traces. It also monitors signal quality, identifying potential interference such as electrical noise, clipping, and excessive movement.
DysphaGrip is a research and monitoring prototype, not a medical diagnostic device.
How We Built It
We combined embedded hardware, biomedical signal processing, and web technologies to create an integrated sensing platform.
An Arduino UNO Q interfaces with the MPU-6050 motion sensor, while a contact microphone captures throat-surface vibrations. A Python processing service synchronizes the signals using a shared timeline.
Our dashboard visualizes the resulting measurements, allowing users to inspect throat vibrations, movement patterns, and signal quality.
We also explored publicly available clinical swallowing data to ground our research. The dataset includes recordings from 248 dysphagia patients and more than 3,100 swallowing events, sampled at 4,000 Hz and associated with clinical imaging verification.
These recordings provide a scientific foundation for future signal-processing research, although they do not independently establish the diagnostic accuracy of our prototype.
Challenges We Faced
One of our biggest challenges was integrating two fundamentally different types of sensors into a synchronized recording system.
Throat vibrations and neck movements occur on different scales and can be affected by background noise, sensor positioning, and unrelated movements.
We also had to consider how to distinguish useful signals from artifacts without making unsupported medical claims.
Building a system that combines hardware acquisition, Python signal processing, and an accessible visualization dashboard within a hackathon timeframe required careful coordination.
Accomplishments We're Proud Of
We developed a wearable sensing approach that brings together acoustic and motion measurements in a single research platform.
Our key accomplishments include:
- Combining throat vibration and six-axis motion sensing.
- Developing a synchronized signal-processing architecture.
- Designing visualizations for waveforms, spectrograms, and motion traces.
- Incorporating automated signal-quality checks.
- Grounding our research in clinical swallowing recordings.
Most importantly, we explored how accessible hardware could support future research into swallowing health.
What We Learned
This project taught us that biomedical sensing is not simply about collecting data. Reliable measurements require synchronization, noise reduction, careful sensor placement, and an understanding of physiological signals.
We also learned the importance of separating technical feasibility from clinical validation. Detecting a vibration or movement pattern does not automatically mean that a system can diagnose dysphagia or identify aspiration.
What's Next for DysphaGrip?
Our long-term vision is to develop DysphaGrip into an affordable, clinically validated swallowing-monitoring platform.
Future improvements include:
- Training machine-learning models to recognize swallowing-related signal patterns.
- Expanding testing across different users and swallowing conditions.
- Improving wearable comfort and sensor placement.
- Conducting controlled evaluations against established clinical assessments.
- Exploring real-time feedback for researchers and healthcare professionals.
Our vision is simple: make swallowing-related research more accessible through affordable hardware, meaningful data, and intuitive technology.
Technologies Used
Arduino UNO Q, MPU-6050, contact microphone, Python, TypeScript, biomedical signal processing, acoustic analysis, accelerometer and gyroscope sensing, clinical swallowing datasets, and web-based data visualization.## Inspiration
Built With
- accelerometer
- arduino-uno-q
- biomedical-engineering
- clinical-data
- contact-microphone
- css
- data-visualization
- dysphagia
- gyroscope
- healthcare
- html
- javascript
- mpu-6050
- node.js
- python
- react
- signal-processing
- typescript
- vite
- wearable-technology
