If you have a blood pressure monitor (a device with an inflatable cuff used to measure blood pressure) and are meticulous about using it, you can measure your own blood pressure many times a day. If not, you probably only get your blood pressure checked when you go to the doctor, maybe a few times a year.
Is it enough? Blood pressure changes over time, sometimes even from minute to minute, not just from year to year or from day to day. It increases when we are busy with activity or is under stress and decreases when we are relaxed. People with "white coat hypertension" who are very worried about doctors can only control their blood pressure when it is abnormally high, which can lead to prescription drugs or other unnecessary treatments.
In the world of smartwatches and fitness trackers that constantly monitor heart rhythm, skin temperature, sleep quality, and more, blood pressure stands out as an important statistic missing from the data set of devices. The blood pressure monitor is too big and bulky.
The new blood pressure sensor was created by Roozbeh Jafari (Texas A&M University), Deji Akinwande (Austin University) and colleagues to close the data gap in blood pressure measurements. The sensor uses a temporary tattoo made of graphene. and is protected by an ultrathin polymer film (bright patches in figure 1) and measures bioimpedance, which is the tissue's resistance to a modified electrical current.
Currently, a complex machine learning system that requires hours of training for each new user must convert bioimpedance to blood pressure. The research team's goal is to turn their sensors into a plug-and-play device as portable as a smartwatch, adding blood pressure to the list of vital signs that are constantly being monitored.
Blood pressure monitors have been around for 100 years. The way it works has not changed significantly since Russian doctor Nikolai Korotkov identified the critical distinction between systolic blood pressure (the maximum pressure at the peak of a blood pulse) and diastolic blood pressure in 1905 (the lowest pressure between blood pulses). The cuff first widens until it completely covers the artery. After that, the cuff is gradually deflated so that blood can pass through part of the heartbeat cycle and eventually the entire cycle. The result is a pair of pressures, expressed in millimeters of mercury, that together form the standard blood pressure measurement.
The sphygmomanometer cuff is bulky and difficult to use because it has to be tightened tightly enough to stop the blood flow. However, Jafari, Akinwande, and colleagues tried to develop a blood pressure meter that does not detect mechanical pressure at all, but detects other associated values. Some emerging technologies can measure blood pressure using smaller, more sensitive pressure sensors.
An expert in machine learning techniques for biomedical applications, Jafari has been working on calculating blood pressure using bioimpedance for several years. Because blood is an ion-rich liquid, it conducts electricity better than most other tissues. It shows how the overall tissue impedance is reduced when the blood pulse passes through an artery. In addition, faster blood pulse spread is associated with higher blood pressure. Therefore, the pulse transit time between two points where bioimpedance is measured on the same artery is another important data.
Yet there is no clear and direct link between impedance and blood pressure. To determine blood pressure, it is necessary to extract sensitive features of the form of the impedance curve throughout the pulse cycle. Machine learning can help with this.
The first bioimpedance measurements were made using standard electrodes similar to those used by Jafari and colleagues to capture electrocardiograms. Attaching the electrodes for long periods of time can be painful, as a strong adhesive is required to hold the electrodes in place. Worse, as the adhesive layer expanded and contracted over time, the placement of the electrodes on the skin would change, resulting in measurement error. This is impractical for an instrument designed for continuous measurement.
Akinwande was also exploring potential uses of graphene-based biomedical sensors. Researchers have been developing tests and technologies based on graphene since two-dimensional carbon material first appeared in the scientific field 20 years ago. “I was just thinking about what opportunities graphene has to offer,” Akinwande said.
Many of these opportunities have emerged in the field of flexible, wearable electronics. Because graphene is a semi-metal, it can be used to create electrodes that can detect electrical signals from the human body or other sources. It is atomically thin and adheres to the skin only with Van Der Waals forces and does not require the use of adhesives.
The entire structure is only a fraction of a micron thick, even with the protective polymer layer needed to prevent the graphene from rubbing just on top of it. The user will not even notice that the graphene is there because it is thin and flexible enough to fit into all the creases and folds of the skin.
Akinwande calls the electrodes "graphene electronic tattoos" because of their similarity to ink-based temporary tattoos. Tattoo paper specially designed to apply the tattoo is applied to the skin and moistened with water. Surprisingly resilient, tattoos can be left in place for up to a week before inevitably peeling away with a layer of dead skin cells. They are unaffected by routine activities or even moderate washing.
Akinwande and colleagues used graphene tattoos on the subject's skin near the eye in one of their tests to analyze the electrical impulses from the subject's eye muscles and detect the subject's gaze with several degrees of accuracy. The signals were used to control a robotic drone that the subject could control by simply looking around the room.
The importance of tattoos in blood pressure measurement is that they do not move over time. Its consistent measurements are perfect for training and subsequent application of Jafari's machine learning algorithm.
The new blood pressure sensor places six consecutive patches of graphene along the radial artery on the side of the wrist closest to the thumb. The opposite ulnar artery is covered by six more that are not visible. Each pair delivers an imperceptibly small electrical current to the wrist through electrodes at both ends. The remaining four are split into two pairs with each pair measuring the induced potential difference, which is inversely proportional to the impedance.
The researchers had a small group of volunteers who wore both graphene tattoos and regular sphygmomanometers for hours while engaging in activities to raise and lower blood pressure to train the machine learning system.
The peak and trough, maximum slope, and pulse transit time of each pulse cycle, among other information, were taken from the bioimpedance curves and fed to the machine learning algorithm by the researchers. When training was complete, the algorithm accurately predicted both systolic and diastolic blood pressure to roughly within 5 mm Hg; this is a great result in the context of blood pressure readings.
It goes without saying that many of the benefits of a light blood pressure measurement are lost if you have to wear a blood pressure monitor for hours to train. The researchers' ultimate goal is to be able to tattoo a graphene material on a fresh person and immediately get an exact reading of their blood pressure.
They're not there yet, but they want to move forward by investigating various amounts of bioimpedance curves that may be less user specific, such as ratios rather than absolute values. A few days after the trial, one of the subjects got a new tattoo, which was a positive result. Without any extra algorithm training, they achieved accurate blood pressure readings within 10 mm Hg: not as good as in the original experiment, but still a usable measurement.
Jafari believes that if continuous blood pressure monitoring becomes common, it could cause a shift in the way people see blood pressure measurement errors. The current paradigm treats each measurement and error bars independently, as readings are rarely made. However, he adds, “absolute error is less important with continuous monitoring.” “The important thing is the trend. Does your blood pressure rise when you are stressed and fall when you lie down? How much longer? ”
Another benefit is that tattoos can be placed almost anywhere on the body. Although the measurement from a sphygmomanometer is typically taken around the brachial artery in the upper arm and is referred to as "your blood pressure", blood pressure varies depending on local factors in different parts of the body, such as stiffness of certain arteries.
Data on blood pressure in various body regions, such as the arteries in the neck that supply blood to the brain, are too risky or impractical to investigate with age-old arterial tightening technology, which can be useful for patients with poor circulation.