Recent research led by Singapore Management University Assistant Professor Dong Ma seeks to unleash the potential of photoplethysmography for wearable healthcare in an era of earbuds, smartwatches, and fitbits.
Image Credit: AK Fotos321/Shutterstock.com
Photoplethysmography (PPG) is a mouthful, even by medical terminology standards. However, it is frequently employed in clinical settings as a noninvasive optical method for determining a patient's vital signs, including blood oxygen saturation and pulse rate. The most common form is a finger-attached clip-on oximeter.
PPG was first created in the 1930s and uses light to illuminate the skin's microvascular bed. After that, a photodiode placed next to the light source records the reflected light, also known as the PPG waveform, which offers information about the cardiovascular system. Oximeters have been a common part of patient monitoring during anesthesia in the United States since the 1980s.
The 2010s saw the emergence of consumer health wearables, including fitness trackers and smartwatches, which are currently in use by millions of people. These devices use PPG, but compared to PPG used in stationary settings, environmental factors restrict the range of accurate information they can provide.
New research led by Dong Ma, an Assistant Professor in Singapore Management University's (SMU) School of Computing and Information Systems, aims to identify ways to unleash PPG's potential for wearable healthcare. The project has received a Tier 2 grant from the Ministry of Education (MOE) Academic Research Fund (AcRF).
I aim to enhance the accuracy and reliability of PPG-based monitoring of human vital signs and unlock new capabilities for human cardiovascular monitoring in real-world scenarios, using wearable devices.
Dong Ma, Assistant Professor, Singapore Management University
Significant Challenges
“The world is facing the grand challenge of an increasingly aging population. Increasing people’s awareness of health management and developing wearable technology to monitor health status in daily life, especially for the elderly, plays a vital role in this campaign,” said Professor Ma.
Any factor that affects the propagation path of light will distort the PPG signal and affect the sensing performance.
Dong Ma, Assistant Professor, Singapore Management University
“Motion artifacts, such as different active movements of the user, can cause displacement between the PPG sensor and human skin. As a result, the transmitted light and received light may not align well with the blood pulses, introducing noise to the recorded PPG signal,” elaborates Professor Ma.
Posture is the second factor that prevents wearables from providing accurate PPG readings.
“When human posture changes during wearing a smartwatch, the contact pressure (tightness) between sensor and skin also changes, resulting in the deformation of skin and the quality of the PPG signal,” added Professor Ma.
It raises the question of why research like Professor Ma's has taken so long to come to fruition, considering that wearable technology is designed to be utilized in active situations and has been integrating PPG since the 2010s. Ma notes that, in reality, over the past few decades, there has been constant research into ways to improve PPG signal quality.
But the [proposed] solutions evolve from pure signal processing to more advanced learning-based approaches. In my opinion, there are two key reasons why improving a distorted PPG signal is challenging: first, the intended signal (blood pulses) is much weaker than external motion artifacts; and second, when there is a displacement (such as sensor detached from the skin), the intended blood pulses are completely absent in the measured signal, making restoration extremely challenging.
Dong Ma, Assistant Professor, Singapore Management University
Pressure Sensors
Incorporating the device's contact pressure (CP) against the skin into the computations is the innovative solution put forth by Professor Ma and colleagues.
“Currently, there is very little research exploring the impact of CP on the PPG signal. And all of them just demonstrate CP can affect the waveform pattern of the PPG signal, but do not propose any solution to address the issue. My proposal aims to integrate CP as a main factor,” said Professor Ma.
“Since the proposal submission, we have started working on the project and have successfully developed an effective transformation model that can reconstruct CP-distorted PPG waveforms. That study is currently under review,” added Professor Ma.
However, a step into the real world will necessitate the creation of transparent pressure sensors that do not block PPG light flows.
“Based on initial exploration, currently there are no commercially available transparent pressure sensors. But during my interaction with researchers from material science, [I have found] it is technically possible to manufacture such pressure sensors,” said Professor Ma.
“And of course, my proposal can promote their development, as material researchers are always looking for a good application for their development. I am seeking collaboration with these researchers to build the sensor,” noted Professor Ma.
Real-World Monitoring
“Current PPG-based wearable devices are good at measuring HR (heart rate) and SpO2 (oxygen saturation level) in both stationary and motion cases. They also support the measurement of HRV (heart rate variability) and BP (blood pressure) but are constrained in stationary cases with a proper setting to obtain a good performance,” said Professor Ma.
“The performance degrades in active scenarios as they require more fine-grained and sophisticated features. Other applications, such as breathing rate estimation, vascular aging, atherosclerosis, and blood glucose estimation have been demonstrated with PPG in [stationary] research, but are still challenging to deploy on wearable devices for real-world monitoring,” Professor Ma noted.
The findings of Professor Ma's research are especially relevant to Singapore.
“First, Singapore is one of the countries that is considered super-aged in the near future. Wearable healthcare technology is an effective weapon in elderly care, promoting an active lifestyle and fostering overall wellbeing,” said Professor Ma.
Professor Ma added, “Second, cardiovascular disease accounted for 31.4 percent of all Singaporean deaths in 2022, and more than 400,000 Singaporeans are living with diabetes. These diseases require continuous and daily monitoring of the health status, whereby PPG-based wearable technology can help.”
“Many of my previous research projects have focused on human behavior sensing, but I found physiological sensing has more significance as it reflects how internally the human body is functioning. And I believe with the increased awareness of health management and global aging, there are more needs and challenges that can be explored in the future,” concludes Professor Ma.