CARDIAC DIAGNOSING
BY A PIEZOELECTRIC-TRANSDUCER-BASED
HEART SOUND MONITOR SYSTEM
Shinichi Sato
1
, Takashi Koyama
2
, Kyoichi Ono
1
, Gen Igarashi
2
, Hiroyuki Watanabe
2
, Hiroshi Ito
2
1
Department of Cell Physiology, Akita University Graduate School of Medicine, 1–1–1 Hondo, Akita, Japan
2
Department of Cardiovascular and Respiratory Medicine, Akita University Graduate School of Medicine
1–1–1 Hondo, Akita, Japan
Mikio Muraoka
Department of Mechanical Engineering, Akita University Graduate School of Engineering and Resource Science
1–1 Gakuen-machi, Tegata, Akita, Japan
Keywords: Piezoelectric transducer (PZT) sensor, Heart sound, Super low frequency sound, Cardiac diagnosis,
Phonocardiographic diagnosis.
Abstract: A variety of diagnosis that auscultation enables is not necessarily conducted thoroughly by all of the
physicians because of its difficulty in judging by means of listening to evanescent sounds with their ears. A
system that displays heart sounds continuously on a computer screen may be convenient for cardiac
diagnosis. Recently, we made a monitor system with employing a piezoelectric transducer (PZT) sensor,
which detected 1
st
and 2
nd
sound and murmurs clearly. In addition, the sensor was capable of detecting
inaudible low-frequency sounds of below ~20Hz. Using the PZT sensor, we recorded heart sounds at left
second intercostals space in 12 patients susceptible to cardiac diseases in parallel with ECG recording, and
analysed the raw heart sounds and band-pass (20–100 Hz) filtered signal. Second sound in the filtered signal
was completely or often absent and/or a sharp deflection, which appears coincidently with R wave, with a
peak-peak voltage of >20 mV and a duration of <25 ms was missing in the raw sound signal in 90% (9/10)
of the patients diagnosed as having cardiac dysfunctions. Thus, we believe that the PZT-based heart sound
monitor system may contribute to the advance of the phonocardiographic diagnosis of cardiac diseases.
1 INTRODUCTION
The stethoscope was invented by Laennec in 1816
and a diaphragm stethoscope, which was invented
by Bowels in 1894, made a great progress in
stethoscope as it increased the sensitivity in mid-
high frequency range to detect heart sounds (1
st
and
2
nd
sound) clearly. However, the diagnosis by
auscultation is not necessarily conducted by all of
the physicians because of its difficulty in judging by
means of listening to evanescent sounds with their
ears. A monitor system that enables us to observe
those heart sounds continuously on a computer
screen may be of help for the phonocardiographic
diagnosis of cardiac dysfunctions.
Monitoring of heart sound using a piezoelectric
transducer (PZT) sensor seems to be an alternative
to the auscultation using a stethoscope. In fact, we
were able to continuously observe the heart sound
signal of an anesthetized small animals on a
computer screen using the PZT sensor (Sato et al.,
2006; Japan patent 4015115; US patent 7174854)
with a custom heart-sound detector circuit (Sato.,
2008). Furthermore, the PZT sensor was capable of
analysing the change in development of cardiac
function during the early postnatal period in mice
(Sato, 2008). Following the animal experiments, we
performed long-term measurements of heart rate of
newborn infants in the neonatal intensive care unit
(NICU) with a PZT sensor that was modified for the
use for the neonates. The study demonstrated that
280
Sato S., Koyama T., Ono K., Igarashi G., Watanabe H., Ito H. and Muraoka M..
CARDIAC DIAGNOSING BY A PIEZOELECTRIC-TRANSDUCER-BASED HEART SOUND MONITOR SYSTEM.
DOI: 10.5220/0003288702800283
In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2011), pages 280-283
ISBN: 978-989-8425-37-9
Copyright
c
2011 SCITEPRESS (Science and Technology Publications, Lda.)
the PZT sensor is quite non-invasive and practically
available as a cardio-respiratory monitor for use in
the NICU (Sato et al., 2010). Recently, we have
made a new heart sound monitor system with
employing a PZT sensor (patent pending), which can
be used as a stethoscope and detected 1
st
and 2
nd
sound and murmurs as clearly as an electronic
stethoscope did. In addition, the sensor was capable
of detecting inaudible low-frequency sounds of
below ~20Hz (super low frequency sound). In the
present preliminary study, we recorded heart sounds
at left second intercostals space in 12 patients
susceptible to cardiac diseases using the PZT sensor
in parallel with ECG recording. Then we analysed
the raw heart sounds and band-pass (20–100 Hz)
filtered heart sound signal to find out the differences
between the patients with and without cardiac
dysfunctions.
2 METHOD
2.1 PZT Heart Sound Monitor System
The PZT heart sound monitor system consists of a
PZT sensor, A/D converter (Power Lab 26T,
ADInstruments) and a computer, which contains a
signal analysis software (Chart5, ADInstruments)
(Fig. 1). The first prototype PZT sensor consists of a
plastic cylinder (20 mm outer diameter, 90 mm long
and 1 mm-thick wall) and a disk-shaped thin
ceramic-PZT (20 mm diameter), which was adhered
on one open edge of the cylinder.
Figure 1: A photograph of the first prototype PZT sensor
(upper) and block diagram of the PZT heart sound monitor
system.
2.2 Output Signal of the PZT Sensor
and an Electronic Stethoscope
Although the PZT sensor detects a wide frequency
range of heart sound from a patient including super
low frequency sound of below ~20Hz, first and
second heart sound signals were clearly separated
from the PZT-sensor output signal by filters with
pass band of 20–100 Hz, which were comparable
with heart sounds obtained by an electronic
stethoscope (Littman Model 3100) from the same
patient as shown in Fig. 2.
Figure 2: Heart sounds obtained by the PZT sensor (upper)
and an electronic stethoscope (lower).
2.3 Cardiac Signal Recording
Twelve outpatients who visited to the cardiology
outpatient department (internal medicine) in the
Akita University Hospital were enrolled in this study.
In parallel with ECG recording, cardiac signal
recording by the PZT-sensor was performed at
second, third, fourth and fifth left-intercostals spaces
(apex region) in addition to the second right-
intercostals space of the patients. Each recording
time at these sites was within 10 seconds and stored
into a computer at a sampling rate of 5 kHz. Sound
filtering condition was set by using the Chart5
software for the separation of 1
st
and 2
nd
sounds
from other lower frequency components.
2.4 Heart Sound Analysis
In the present study, the analysis of heart sounds was
performed with the signals obtained at the left
second intercostals spaces of the patients. The raw
heart sound signal was analyzed with focusing on
the period of the QRS wave of simultaneously
recorded ECG. A sharp deflection, which appeared
coincidently with R wave, was measured by its
peak-to-peak voltage (Vp-p, mV) and its rising or
falling time (ms: duration). While, the 1
st
and 2
nd
CARDIAC DIAGNOSING BY A PIEZOELECTRIC-TRANSDUCER-BASED HEART SOUND MONITOR SYSTEM
281
sounds in the band-pass filtered PZT sensor signal
were qualitatively scored as “Y” (perceivable) or
“N” (non-perceivable). Regardless of the sound
analysis, cardiac diagnosis of the patients was
performed in a conventional way by a physician in
the cardiology outpatient department.
3 RESULT
According to the result of diagnoses conducted by
the physician in the cardiology department, 10 out of
total 12 patients had following one or several cardiac
dysfunctions; aortic valve regurgitation (AR), mitral
valve regurgitation (MR), pulmonary valve
regurgitation (PR), Aortic valve sclerosis (AVS), or
dilated cardiomyopathy (DCM). In the heart sound
analysis, second sound in the filtered signal [arrows
at middle trace in Fig. 3(a)] was completely or often
absent in 6 out of the 10 patients who had above
described cardiac dysfunctions [middle trace in Figs
3(b) and (d)].
Table 1: Results of cardiac diagnosis and sound analysis.
patient heart disease 2
nd
sound Vp-p (mV)
1 normal Y 58
2 AR, MR, PR Y -
3 AR, MR, PR N -
4 DCM N -
5
AR
N 27.6
6 MR, PR Y 10.5
7 MR, PR, DC
M
Y -
8 normal Y 25.5
9 AVS N 22.5
10 AR, MR, PR N -
11 MR Y 58.4
12 AR N -
Furthermore, a sharp deflection [arrows at top trace
in Fig. 3(a)], which appears coincidently with R
wave in healthier patients, with an average of three
consecutive peak-to-peak voltages (Vp-p) of >20
mV and a duration of <25 ms was missing in the raw
sound signal in 7 out of the 10 cardiac patients [top
trace in Figs 3(b) and (d), Table 1]. Accordingly,
patients who had at least either one of above 2 heart
sound deficits were 90% (9/10) of the patients
diagnosed as having cardiac dysfunctions. Two other
patients were scored as normal with cardiac function
by both the sound analysis and a diagnosis
conducted by the physician.
Figure 3: From top to bottom trace; raw heart sounds,
band-pass filtered heart sounds, and ECG of (a) patient #1
with normal cardiac function and patients (b) #3, (c) #6,
(d) #10 with cardiac dysfunctions as shown in Table 1.
4 DISCUSSIONS
Since two decades ago, many researchers argued the
possibility in analysing cardiac function by using
heart sounds or phonocardiogram signal processing
technique (Rangayyan and Lehner, 1988; Durand
and Pibarot, 1995; Manecke et al., 1999). Electronic
stethoscope, though it is cutting edge equipment and
commercially available these days, it seems that it
still not has an overwhelming advantage against the
conventional stethoscope.
BIODEVICES 2011 - International Conference on Biomedical Electronics and Devices
282
The striking difference between our PZT heart
sound monitor system and the electronic stethoscope
is the sound detection element; a small vibration
detecting PZT disk adhered on the top of a plastic
cylinder (Fig. 1) directly touch a patient’s chest and
record heart sounds, while heart sounds are detected
by a microphone via the air inside the head of the
electronic stethoscope. The construction of the PZT
sensor is crucial for the detection of super low
frequency sound of below ~20Hz, which may be out
of range for an electronic stethoscope because its
sensor head is diaphragm type that intends to detect
mid-high frequency range sounds.
In the present study, we analysed the raw heart
sound signal that contained the super low frequency
sound, and we found that the raw sound signal of
cardiac patients lacked a sharp deflection that
appears in coincidently with R waves on ECG. We
also found that 2
nd
sound in the filtered heart sound
was missing in most of the cardiac patients.
However, the 2
nd
sounds were often observed in the
heart sound signal recorded at 3
rd
, 4
th
or 5
th
(apex)
intercostals spaces. Accordingly, the recording of
heart sound signal with the PZT sensor at left 2
nd
intercostals space seems to be effective for the
diagnosis of patients susceptible to cardiac diseases.
It should be noted that the analysis of the super-low-
frequency sound is likely to be useful for cardiac
diagnosis (see Fig. 3).
As we only found a small part of information
hidden in the low frequency heart sound, we need
further investigate the sounds created by the heart in
cardiac patients with developing a quantitative
method for the cardiac diagnosis including such as
frequency domain analysis. Furthermore, we should
develop a simple and easy to use heart sound
monitor system, which continuously displays the
cardiac signal of a patient and provides us a tool for
visually diagnosing for the use in cardiology
outpatient department in hospitals.
In conclusion, the present study demonstrated
that the PZT-based heart sound monitor system has a
performance suitable for detecting heart sounds
including super low frequency sound. Although our
results are very preliminary and we may need to do a
further comparative analysis with other electronic
stethoscopes, we believe that phonocardiogram-
based analysis with the PZT heart sound monitor
system may provide us a new strategy for the
diagnosis of cardiac diseases.
ACKNOWLEDGEMENTS
This work was supported in part by the Vehicle
Racing Commemorative Foundation, Suzuken
Memorial Foundation, Nakatani Foundation of
Electronic Measuring Technology Advancement and
an intramural grant from Akita University. The
authors thank the staff of the department of internal
medicine at Akita University Hospital and Yu Obara,
Yuta Nakamura and Hideaki Kobayashi for their
contributions to the examination of the prototype
PZT sensor.
REFERENCES
Durand, L. G., Pibarot, P., 1995. “Digital signal
processing of the phonocardiogram: review of the
most recent advancements”, CRC Crit Rev Biomed
Eng 23: 163–219
Manecke, G. R. Jr., Nemirov, M. A., Bicker, A. A.,
Adsumelli, R. N., Poppers, P. J., 1999. “The effect of
halothane on the amplitude and frequency
characteristics of heart sounds in children”, Anesth
Analg 88: 263–267.
Rangayyan, R. M., Lehner, R. J., 1988. “Phonocardiogram
signal analysis: a review”, CRC Crit Rev Biomed Eng
15: 211–236.
Sato, S., K, Yamada., N, Inagaki., 2006. “System for
simultaneously monitoring heart and breathing rate in
mice using a piezoelectric transducer”, Med Biol Eng
Comput 44 (5): 353–362.
Sato, S., 2008. “CARDIAC BEAT DETECTOR”,
BIOSIGNALS 2008 Proceedings 2:136–140.
Sato, S., 2008. “Quantitative evaluation of ontogenetic
change in heart rate and its autonomic regulation in
newborn mice with the use of a noninvasive
piezoelectric sensor.”, Am J Physiol Heart Circ
Physiol 294: H1708–H1715.
Sato, S., Nakajima, W., Ishida, A., Kawamura, N., Miura,
S., Ono, K., Inagaki, N., Takada, G., Takahashi, T.,
2010. “Assessment of a new piezoelectric-transducer
(PZT) sensor for noninvasive cardiorespiratory
monitoring of newborn infants in the NICU.”,
Neonatology 98:179–190.
CARDIAC DIAGNOSING BY A PIEZOELECTRIC-TRANSDUCER-BASED HEART SOUND MONITOR SYSTEM
283