Abstract
Vital health parameters can provide a staggering amount of information about the general health of an individual, and continuous vital monitoring plays a significant role in the identification of clinical deterioration. Ballistocardiography (BCG), a technique to measure the micro-vibrations caused by cardiac contractions and body movements, allows for long-term non-invasive, contactless vitals monitoring. However, the effect of external factors, such as body posture, sensor position, bed angle, and mattress thickness, on the BCG signals remains to be analyzed in detail. Hence, this research focuses characterizing BCG signals concerning the aforementioned factors by employing Dozee, a BCG-based contactless continuous vital parameters monitoring system, in a simulated use environment. Data was gathered from 10 volunteers, and the waveforms for HR and RR were extracted. Unsupervised learning models were used to analyze the signal quality and observe the trends among the volunteers to gauge the impact of the studied factors. The optimal signal was observed across individuals in the supine position, with the sensors positioned near the head and heart; bed angles between 30° and 50°, and mattress thicknesses between 5.5-8.5 cm. This study will aid in improving the signal-to-noise ratio of the BCG signals and perhaps identify the external factors influencing it, contributing to improved non-intrusive an monitoring experience. non-intrusive patient
Introduction
Patients' vital signs like Heart Rate (HR) and Respiratory Rate (RR) are monitored throughout clinical care in hospitals and at home. This is because changes to these parameters are a good indicator of health and well- being. Continuous vital sign monitoring in these situations is often used for early identification of clinical deterioration to ensure appropriate actions are taken. Shifts in the patterns and absolute values of vital signs often precede emergency events like cardiac arrest, stroke, [1][2] or subsequent death. Thus, continuous monitoring of vital signs may be a useful tool to detect clinical deterioration at an earlier stage and enable clinicians to take corrective interventions. Technology involving continuous vital sign monitoring has advanced significantly in recent years. One such breakthrough is the analysis of micro-vibrations using Ballistocardiography (BCG) based devices for continuous vitals monitoring. Continuous monitoring using BCG has also proven to be a useful tool for analyzing sleep disorders [3]. In essence, Ballistocardiography is a technique that graphically represent the heart's repetitive internal movements caused by the inflow of blood. During atrial systole, when blood is pumped into the major vessels, the center of mass of the body shifts towards the head, but during diastole, blood travels into the peripheral arteries of the body, changing the center of mass towards the feet. Because of changes in blood distribution during the cardiac cycle, the BCG waveform is affected by a shift in the body's center of mass [4]. These waveforms can be studied to assess cardiac and respiratory cycles. Traditionally, Electrocardiography (ECG), which works on the principle of monitoring the heart's electrical activity, has been used for continuous cardiac monitoring. The major difference between ECG and BCG is that while ECG measures cardiac electrical activity, BCG measures the mechanical vibrations which are caused by electric signals. The ECG QRS complex shows the electrical signals, i.e., ventricular depolarization before the contraction [5], whereas the IJK peak in BCG occurs due to the physical movement during the ventricular contraction [6], and therefore there is a time lag between them. (Fig. 1)
Despite its advantages, BCG and similar techniques have had trouble finding practical applications because of two main issues: (a) it picks up forces around the sensor, which can include human body movements, respiration, and snoring, which affect the accuracy and detection rate of cardiac and respiratory measurements; and (b) the measurements can change depending on the setup and placement of the sensors, as well as body postures. Dozee, a device for monitoring vital signs based on BCG, has been developed by Turtle Shell Technologies Pvt. Ltd. With a novel unsupervised clustering method that successfully separates cardiac contractions and respiratory events from the unconstrained BCG signal, the Dozee device has overcome the two major issues mentioned above. Dozee is intended to serve as a remote patient monitoring system for tracking vital signs and sleep. The patient, doctor, caregiver, or user can access the acquired data through a mobile application or a web-based platform (Fig. 2) [7]. This device's precision of HR and RR measurements has been evaluated against an FDA- approved patient monitoring system which has been published elsewhere. [7]
It is important to note that the BCG signals are strongly influenced by the measurement tool, subject's postures, medium for signal transmission, etc. BCG signal processing is therefore an extremely challenging task for researchers. Additionally, several factors, e.g. mattress thickness, body movements, motion artifacts, etc., can affect signal properties [8].
The raw signal is noisy and nonstationary due to body movement, induced respiratory efforts, and the characteristics of the sensing system itself. Moreover, the vibrations from the device are recorded at a rate of 2000 Hz and the human resting heart rate is generally between 60 and 100 beats per minute, which is equivalent to 1Hz- 1.6 Hz and the normal respiration rates for an adult person at rest range from 12 to 20 breaths per minute, equivalent to 0.05-.0.08 Hz. Hence, it is important to eliminate motion artifacts, interferences like AC noise and snoring, and other sources that can degrade the signal's quality. Therefore, characterization of the fluctuations caused in the BCG signal as a result of these variables can assist in a better understanding of the nature of the artifact impacting the data and can be taken into account when analyzing the signals for vital signs.