About PPG, ECG, and why we are interested in their relations?

The heart is activated by an electrical stimulus produced by the sinoatrial (SA) node, located in the atria of the heart. This causes muscle contraction, initiating the process of blood being pumped to the rest of the body. The electric field produced by the heart causes current to flow in rhythm with the heart’s cardiac cycle. ECG monitors the heart’s cardiac activity by measuring this electrical activity.

PPG signals are recorded using optical methods that monitor the blood volume in the skin’s tissue. As the heart contracts, blood flows to the extremities of the body. A photodiode attached to a particular extremity, like the finger or ear, can capture the changes in blood volume as the intensity of reflected light emitted from the diode is modulated according to the blood volume at the measurement site.

Because the blood volume changes at the extremities of the body are caused by the contractions that are induced by the SA node, the ECG and PPG signals are intrinsically related. By exploiting this relationship, it becomes possible to synergistically leverage the rich clinical ECG knowledge and powerful real-time monitoring capabilities of PPG to enable low cost, everyday cardiac monitoring.

Understanding the PPG-ECG relationship

In 2019 approximately 55 million mortalities worldwide were associated with cardiovascular (CV) disease (CVD). This severe illness is the leading cause of mortality worldwide, exceeding the second leading cause–stroke–by approximately two million deaths annually¹. Early treatment of CVDs can effectively reduce the risk of sudden death. While, the presence of clinical symptoms usually indicates the onset of underlying cardiac dysfunction, certain cardiac issues show no obvious clinical symptoms in the early stage. The lack of clinical awareness often leads to a delay in patients’ early intervention opportunities, which disproportionately affects low-income, disadvantaged populations.

The gold standard for non-invasive cardiac monitoring and diagnosis of CVDs is an electrocardiogram (ECG). Long-term continuous ECG monitoring can help in capturing the intermittent and asymptomatic abnormalities of the heart. While ECG monitoring is available in smartwatches and smartphones, these products require consistent user engagement, making long-term continuous monitoring infeasible. The goal of this study is twofold. First, we aim to develop artificial-intelligence-driven (AI-driven) health solutions for automated and continuous cardiac monitoring by inferring ECG signals from photoplethysmogram (PPG) signals, which can be monitored continuously without constant user attention. Second, we aim to jointly analyzing the ECG-PPG relationship to help us discover more about the underlying mechanisms that govern the structure of each signal. Our work that has been devoted to examining these issues can be found in the publications cited below.

Developing a physiological digital twin framework

Understanding relationship between electrocardiogram (ECG) and photoplethysmogram (PPG) signals has the potential to enable long term continuous monitoring. Realizing the full potential of such technology relies on how the technology is deployed under different use cases. The digital twin framework presents an opportunity for realizing precision health by providing fine-grained analysis of individual’s health status. A digital twin is a digital representation of an object, which can be used to monitor its status. The healthcare digital twin has the potential to produce fine-grained personalized models for monitoring an individual’s health.

In this project, our goal is to determine how we can formulate and deploy AI-driven physiological digital twin frameworks to generate high-precision personalized health monitoring under realistic situations. We particularly focused on leveraging the continuous PPG monitoring power to build a digital twin model that establishes a patient’s personalized PPG and ECG relationship when only sporadic amounts of ECG sensing data are available. We hope to continue to further this line of research to further understand how different individuals’ ECG-PPG relationships may vary under different environmental conditions.

Join our studies -- Together, we can advance science and technology!

See more information on the tab "Join Our Studies" (under "More").

1. https://www.who.int/news-room/fact-sheets/detail/the-top-10-causes-of-death

Talks & Tutorials

Tutorial at WIFS'21 on Physiological Forensics

Dr. Wu's Talk in Spring 2021 for the UMD Faraway Fourier Talk (FFT) series

Peer-Reviewed Publications: (**to update)

[1] Qiang Zhu, Xin Tian, Chau-Wai Wong, and Min Wu, “Learning Your Heart Actions From Pulse: ECG Waveform Reconstruction From PPG", IEEE Internet of Things Journal, vol. 8, no. 23, pp. 16734-16748, 2021. (See also an earlier preliminary work at BHI'19, [IEEE Xplore] [Slides])

[2] Xin Tian, Qiang Zhu, Yuenan Li, and Min Wu, “Cross-domain Joint Dictionary Learning for ECG Reconstruction From PPG”, IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP'20), Barcelona, Spain, May 2020 (selected as lecture session talk). [IEEE Xplore][Slides]. See also 3-minute Thesis -- 3MT presentation by Xin Tian.

Additional publications/preprints:

X. Tian, Q. Zhu, Y. Li, and M. Wu: “Cross-domain Joint Dictionary Learning for ECG Inference from PPG,” submitted for journal publication, under review. Preprint is available through <http://arxiv.org/abs/2101.02362>.

X. Tian: Ph.D. Dissertation, Univ. Maryland, College Park, Aug. 2022.

Y. Li, X. Tian, Q. Zhu, and M. Wu: “A Lightweight Neural Network for Inferring ECG and Diagnosing Cardiovascular Diseases from PPG,” submitted for journal publication, under revision. Preprint is available through <https://arxiv.org/abs/2012.04949>.

Patent Applications:

Q. Zhu, X. Tian, C-W. Wong, and M. Wu: “Reconstruction of ECG from PPG Signals for Continuous Monitoring and Analytics,” patent filing March 2020.

Funding Acknowledgement:

We gratefully acknowledge the funding support on various aspects this exciting line of research: NSF Smart Health.