The monitoring of heart activity in the form of heart rate (HR) is one of the most routine
measurements conducted in order to determine the basic medical condition of an individual. The
standard average HR of a healthy adult individual is between 60 and 100 beats per minute and
could be lower for individuals in good physical condition. A noticeable deviation from these
standard values indicates a possibility of cardiovascular diseases. In the last couple of decades
heart rate variability (HRV) has been recognized as an important quantitative marker of the
relationship between the autonomic nervous system and cardiovascular mortality, including
sudden cardiac death. The apparent easy derivation by performing mathematical analysis from the
HR has popularized its use. However, the significance and meaning of many different measures of
HRV are more complex than generally appreciated and there is a potential for incorrect conclusions
and for excessive extrapolations. Because of this, the reliability of HR and HRV measuring device
is of utmost importance.
HR and HRV are most frequently measured and observed in clinical applications. Because of their
reliability and easy implementation, electrocardiography (ECG) and optical plethysmography
(PPG) are considered as the “gold standard”. Both are performed in a contact manner, which is
why they are less appropriate for specific groups of patients, e.g. patients with burns, neonates,
etc. Next to this, there is more and more interest of continuous and long-term measurements also
in predominantly economic areas. The mentioned standard methods operate with sensors, which
need to be in direct contact with human skin and therefore limit the everyday or occupational
activities. Furthermore, any possible unwanted movements of the sensors’ location on the skin
may result in noise or other signal deformation, which could make it inappropriate for analysis. As
a result, the need for non-contact measurement process has emerged. The goal of this dissertation
is to present a new non-contact method for measuring of the HR and the HRV. The dissertation is
focused exclusively on the measurement problematics in the engineering and scientific research
sense. Medical aspects are taken into account, provided by the experts in the field of medical
science.
Within this thesis we describe a proposal of non-contact ultrasound measuring of the HR and the
HRV. The method is based on the Doppler phenomenon as it measures the difference in phase
between the transmitted and deducted signal on the selected region of interest, for example on a
person’s neck. The difference of the phase between the two signals is caused by the pulsation of
the carotid artery. The novelty of our method, compared to other studies, is transmitting at two
different ultrasound frequencies at the same time. With the option of selecting the optimal signal
in real time at either frequency, the reliability of the method and the accuracy of the results have
been significantly improved due to the elimination of the standing waves and other possible
environmental disturbances. The measuring device’s signal was captured with an A/D converter
device, whereas the excitation signal was generated with a D/A converter device. Signal processing
was carried out in a dedicated LabVIEW application, which in addition to monitoring in real-time
also enables the recording of a signal for later processing. The results from the experimental
method were directly compared to the reference contact ECG method, obtained with a Biopac
system MP150 with the ECG100C module. The proposed method was tested in three phases. In the first phase, we simulated the contraction
of a vein by measuring the pulsation of a loudspeaker membrane, which was directly connected to
a signal generator. First, we measured the signal at a constant frequency and later the frequency
was raised in a non-specific time interval. In the second phase, we performed measurements on
volunteers who were lying in bed in laboratory environment. This phase was performed by
measuring at different HR. Firstly at lower HR during rest and secondly at elevated HR after 1
minute of exercise. The last phase of the study was carried out in a realistic clinical setting at the
Clinical department of cardiology, University Medical Centre of Ljubljana (UKCLJ). The noncontact
measuring method was performed on a group of patients with various cardiovascular
diseases. For all measurements in the laboratory and clinical environment, HRV time domain
statistical parameters were calculated.
Prior to calculating the statistical parameters, the parts of the signal with known noise artefact were
removed from the recorded signal (for instance individual’s movement, swallowing of saliva, etc.).
The calculated mean value and standard deviation of the differences between the current HR of
the experimental signal at the optimal frequency and the reference signal on the one hand, and on
the other hand the values were obtained based on moving average of the non-contact signal,
independently of the ECG signal. During the standstill phase, in the first case we obtained results
for the group of volunteers in the laboratory environment ΔHR = 0,23 min-1 ± 0,01 min-1, and in
the second case ΔHR = 0,31 min-1 ± 0,88 min-1. After 1 minute of physical activity the following
values were calculated for the first case ΔHR = 0,29 min-1 ± 0,67 min-1 and ΔHR = 0,42 min-1 ±
0,75 min-1 for the second case. The result of the clinical trial described in this dissertation was
ΔHR = -0,41 min-1 ± 1,97 min-1 when compared to the reference method, and ΔHR = 0,50 min-1 ±
2,00 min-1 when calculated based on the moving average. For the HRV parameters in the time
domain, we calculated the relative error between the reference and the experimental signal. For
the group within the laboratory setting, the results of measurements during standstill phase were
AVNN = 0,07 % ± 0,13 %, SDNN = 1,92 % ± 3,54 %, rMSSD = 9,05 % ± 13,93 %, pNN20 = 5,16
% ± 3,03 %, pNN50 = 3,09 % ± 3,53 %. After physical activity, the results were as follows: AVNN
= 0,06 % ± 0,05 %, SDNN = 0,88 % ± 0,48 %, rMSSD = 11,44 % ± 16,66 %, pNN20 = 25,09 % ±
30,86 %, pNN50 = 1,54 % ± 0,86 %. For the given example of measurements in the clinical
experiment, the HRV parameters were AVNN = 0,05 %, SDNN = 32,82 %, rMSSD = 4,04 %,
pNN20 = 8,94 %, pNN50 = 6,32 %.
The pilot research on a group of volunteers in the laboratory and in the clinical setting suggests
that the non-contact measurement of physiological parameters with the proposed method is indeed
possible. Moreover, the results of our study show that the proposed non-contact measurement
method is reproducible and comparable to standard contact measurement methods.
|
