Related Experiment Video
Updated: May 10, 2025
10:17
Real-Time Cardiac Mapping with a Noninvasive Imageless Electrocardiographic Imaging System
Published on: April 11, 2025
179
AI-Powered Noninvasive Electrocardiographic Imaging Using the Priori-to-Attention Network (P2AN) for Wearable Health
Shijie He1, Hanrui Dong1, Xianbin Zhang1
1School of Biomedical Engineering, Sun Yat-sen University, Shenzhen 518107, China.
Sensors (Basel, Switzerland)
|April 28, 2025
Summary
A new AI network, Priori-to-Attention Network (P2AN), stabilizes electrocardiographic imaging (ECGI) for noninvasive cardiac monitoring. This technology improves real-time cardiovascular health assessment using wearable devices, even in noisy conditions.
Area of Science:
- Biomedical Engineering
- Artificial Intelligence in Healthcare
- Cardiovascular Monitoring
Background:
- Smart wearable devices enable continuous, noninvasive health monitoring via biosignal collection.
- Electrocardiographic imaging (ECGI) reconstructs cardiac electrical activity from body surface potentials but faces instability issues.
- The ECG inverse problem's inherent instability limits the practical application of ECGI.
Purpose of the Study:
- To introduce a novel Priori-to-Attention Network (P2AN) to enhance the stability and accuracy of ECGI solutions.
- To develop a method for noninvasive cardiovascular monitoring using AI-powered wearable devices.
- To improve the diagnosis of cardiac conditions like myocardial ischemia and ventricular hypertrophy.
Main Methods:
- Developed a Priori-to-Attention Network (P2AN) integrating physiological knowledge via cross-attention mechanisms.
- Utilized small-scale convolutions for attention computation, leveraging signal properties and electrical propagation.
- Implemented normalization constraints to improve solution accuracy without requiring clinical transmembrane potential (TMP) measurements.
Main Results:
- P2AN significantly improved transmembrane potential (TMP) reconstruction and lesion localization for diagnosing myocardial ischemia and ventricular hypertrophy.
- The method demonstrated high robustness in noisy environments, suitable for wearable electrocardiographic clothing.
- Achieved enhanced spatiotemporal accuracy and noise resilience in ECGI.
Conclusions:
- P2AN offers a stable and accurate solution for the ECG inverse problem, overcoming previous limitations.
- The developed AI network is highly suitable for real-time, noninvasive cardiovascular monitoring with wearable devices.
- P2AN represents a significant advancement in AI-powered healthcare, enhancing diagnostic capabilities for cardiac conditions.
