Related Experiment Video
Updated: May 28, 2026
05:51
Measuring the Carotid to Femoral Pulse Wave Velocity (Cf-PWV) to Evaluate Arterial Stiffness
Published on: May 3, 2018
Automated Detection of Carotid Artery Stenosis Using a Sensitive Accelerometer Wearable Sensor and Interpretable
Houriyeh Majditehran1, Brian Sang1, Nia Desai1
1School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA.
Biosensors
|May 26, 2026
Summary
A new wearable sensor detects carotid artery disease using subtle vibrations, offering a non-invasive screening tool. This technology aids in early detection of stroke risk factors like stenosis.
Area of Science:
- Biomedical Engineering
- Medical Diagnostics
- Wearable Technology
Background:
- Carotid artery disease significantly elevates ischemic stroke risk.
- Current screening methods are limited by accessibility and scalability.
- There is a need for non-invasive, user-friendly diagnostic tools for early detection.
Purpose of the Study:
- To develop and validate a wearable MEMS accelerometer patch for non-invasive carotid artery disease screening.
- To identify and extract interpretable biomarkers from mechano-acoustic vibrations.
- To assess the performance of a machine learning model for classifying carotid artery pathology.
Main Methods:
- Utilized a wearable MEMS accelerometer patch to capture carotid blood flow vibrations.
- Applied Continuous Wavelet Transform (CWT) for time-frequency analysis.
- Extracted spectral and scalogram-derived features and selected six non-redundant biomarkers.
- Employed SHapley Additive exPlanations (SHAP) for model interpretability.
- Validated the approach in a carotid flow phantom and a clinical study of 74 patients.
Main Results:
- Achieved an Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.97 and Area Under the Precision-Recall Curve (AUPR) of 0.947.
- Demonstrated high diagnostic performance with 81.7% sensitivity and 93.6% specificity.
- Achieved 89.8% accuracy, 85.4% precision, and an F1 score of 83.5%.
Conclusions:
- Wearable seismic sensing combined with interpretable machine learning shows significant potential for carotid artery disease screening.
- The developed approach enables fast screening and longitudinal monitoring of carotid arteries.
- This technology can overcome barriers to frequent monitoring and scalable deployment of diagnostic tools.