P. PE^LIN et al.: AUSCULTATION OF A HEART AND VASCULAR ACTIVITY DURING AURICULAR NERVE STIMULATION
91–100
AUSCULTATION OF A HEART AND VASCULAR ACTIVITY
DURING AURICULAR NERVE STIMULATION
PRISLU[KOVANJE AKTIVNOSTI SRCA IN O@ILJA MED
STIMULACIJO AVRIKULARNEGA @IVCA
Polona Pe~lin
1
, Janez Rozman
2,3,*
, Renata Jane`
3
, Nu{a Baj`elj
4
, Samo Ribari~
3
,
Tomislav Mirkovi}
5
1
Division of Gynaecology and Obstetrics, University Medical Centre Ljubljana, [lajmerjeva 3, 1000 Ljubljana, Slovenia
2
Center for Implantable Technology and Sensors, ITIS d.o.o. Ljubljana, Lepi pot 11, 1000 Ljubljana, Slovenia
3
Institute of Pathophysiology, Faculty of Medicine, University of Ljubljana, Zalo{ka 4, 1000 Ljubljana, Slovenia
4
Faculty of Sport, University of Ljubljana, Gortanova 22, 1000 Ljubljana, Slovenia
5
University Medical Centre Ljubljana, Department of Anaesthesiology and Surgical Intensive Therapy, Zalo{ka 2, 1000 Ljubljana, Slovenia
Prejem rokopisa – received: 2022-11-29; sprejem za objavo – accepted for publication: 2023-01-10
doi:10.17222/mit.2022.700
The principal objective of the research was to develop and test a multichannel system intended to capture heart (PCG) and
Korotkoff sound (KS) signals in a human model of transcutaneous auricular nerve stimulation (tANS). In particular, the purpose
was to develop contact auscultation transducers (transducers) capable of capturing PCG and KS at standard auscultation posi-
tions without and during the selective tANS. The scope was to develop transducers capable of capturing body sounds lying in a
portion of the frequency spectrum between 20 Hz and 200 Hz. They should be as insensitive as possible to external sound or
ambient noise and should be capable of compensating for friction noise due to body movements. The system developed con-
sisted of five transducers, where four of them were intended for the auscultation of PCG generated by heart valves, while one of
them was intended to capture KS. The functionality of the transducers was tested in the female model by applying four of trans-
ducers to the standard positions on the chest and one over the Brachial artery. The results show that the system developed was
highly sensitive to PCG and KS, and less sensitive to external ambient sounds and friction noise. Namely, the S1 and S2
heart-sound peaks are present clearly in the recorded PCG signal as well as during the tANS. It was also shown that during the
tANS, heart cycles became slightly shorter and, thus, the heart rhythm slightly higher. Finally, during the tANS, both heart
sounds S1 and S2 became louder. In conclusion, the findings provide some evidence that the sounds captured by the transducers
emanate from the organs under study and are related to their activity and tANS.
Keywords: heart sounds, Korotkoff sounds, auricular nerve stimulation, auscultation transducer
Osnovni namen dela je bil izdelati in preizkusit ve~kanalni sistem namenjen zajemnaju zvo~nih signalov srca (PCG) kakor tudi
zajemanju zvo~nih signalov Korotkoff-a (KS) na ~love{kem modelu skoziko`ne stimulacije avrikularnega `ivca (tANS). Bolj
natan~no, namen je bil izdelati kontaktne pretvornike za prislu{kovanje, ki bi bili sposobni zajeti PCG in KS na standardnih
prislukovalnih mestih brez in med selektivno tANS. @elja je bila izdelati pretvornike sposobne zajeti telesne zvoke, ki se
nahajajo v delu frekven~nega spektra med 20 Hz in 200 Hz. Pretvorniki naj bi bili kolikor je mogo~e neob~utljivi na zvoke, ki so
prihajali iz okolice in zvoke, ki so nastajali zaradi trenja pri gibanju telesa. Izdelani sistem je vseboval pet pretvornikov pri
~emer so bili {tirje namenjeni prislu{kovanju PCG, ki ga ustvarjajo zaklopke medtem, ko je bil eden namenjen zajemanju KS.
Delovanje pretvornikov je bilo testirano na `enskem modelu pri katerem so bili {tirje pretvorniki name{~eni na standardna mesta
na prsih in eden na mesto nad brahialno arterijo. Rezultati ka`ejo, da je izdelani sistem visoko ob~ultljiv na PCG in nizko
ob~utljiv na zvoke, ki so prihajali iz okolice in zvoke, ki so nastajali zaradi trenja. V posnetem PCG signalu so bili namre~
vrhovi zvokov srca S1 in S2 jasno razvidni tako brez kakor tudi s tANS. Pokazalo se je tudi, da so bili cikli srca med tANS malo
kraj{i in s tem ritem srca malo vi{ji. Nazadnje se je pokazalo, da so postali zvoki srca S1 in S2 med tANS glasnej{i. Zaklju~iti je
mogo~e, da dognanja podajajo nekatere dokaze pri katerih so bili zvoki, ki so prihajali iz preiskovanih organov, povezani tako z
njihovo aktivnostjo kakor tudi s tANS.
Klju~ne beside: zvoki srca, zvoki Korotkoffa-a, stimulacija avrikularnega `ivca, prislu{kovalni pretvornik
1 INTRODUCTION
Biomedical signals, such as an electrocardiogram
(ECG), photoplethysmogram (PPG), phonogastrogram
(PGG) and heart sounds or phonocardiogram (PCG), are
frequently used by clinicians to screen and diagnose var-
ious cardiac abnormalities. Accordingly, the analysis of
these biomedical signals has been extensively investi-
gated in the past.
1
Body sounds that are generated by
complex physiological processes within the human body
contain invaluable information about human
physiological and psychological conditions. Each bio-
medical signal, however, has a proprietary set of features
and contains events that can describe corresponding
physiological events.
The non-invasive capture of biomedical signals is rel-
atively simple and inexpensive, if compared to invasive
alternatives. However, to enable the analysis of a particu-
lar waveform in more detail, as many as possible events
in captured biomedical signals should be detected. Un-
fortunately, body sounds are mostly barely audible. The
contact auscultation transducers that are most frequently
used in the capture of biomedical signals are micro-
Materiali in tehnologije / Materials and technology 57 (2023) 1, 91–100 91
UDK 537.8:616.12 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 57(1)91(2023)
*Corresponding author's e-mail:
jnzrzmn6@gmail.com (Janez Rozman)

phones, piezoelectric sensors, and capacity-type sensors.
Most of them are capable of capturing body sounds that
are mainly located within the frequency spectrum, rang-
ing from 20 Hz to 1300 Hz.
2
In particular, a human heartbeat is one of the subtler
body sounds, with a low magnitude from 20 Hz to
200 Hz. PCG occurs during the heart cycle, while blood
flowing turbulently through the heart chambers and heart
valves generates vibrations of composing structures, par-
ticularly when the heart valves open and close. To cap-
ture and analyse the human heartbeat, however, the trans-
ducer should be developed by considering certain
requirements: it should capture heart sounds lying in the
frequency spectrum between 20 Hz and 200 Hz, it
should be as insensitive as possible to external sounds
and noises, and should be capable of compensating fric-
tion noise due to body movements.
3,4
In acoustic stethoscopes, the bell effectively transmits
lower-frequency sounds, while the diaphragm effectively
transmits higher-frequency sounds. In practice, however,
stethoscopes combine these modes into a single mode in-
dependent of the pressure of the stethoscope against the
skin. More descriptively, a higher pressure enhances the
diaphragm mode and higher frequency sounds, while a
lower pressure enhances the bell mode and lower fre-
quency sounds. Electronic stethoscopes, however, enable
the recording of PCG for review and analysis in bell or
diaphragm mode, but are sensitive to ambient and fric-
tion noise. Auscultation of a patient’s PCG is therefore a
safe, inexpensive and very useful diagnostic tool.
The assessment of changes in heart function dynam-
ics is one of the basic procedures used to identify poten-
tially threatening conditions. One low-cost and non-inva-
sive procedure widely used to describe the heart-function
dynamics that are based on the auscultation of its sounds
by placing a stethoscope on the chest is PCG. It is a
graphical representation of heart sounds providing useful
information about the functionality and condition of the
heart valves. Besides, the waveform of the PCG can de-
liver significant information useful to reveal abnormali-
ties in the movement of the heart wall, the closure of
valves, or the leakage of blood flow.
5,6
All of the afore-
mentioned activities generate vibrations that propagate to
the surface of the thorax, where the resulting sounds can
be measured.
7
In this regard, the recorded PCG can be utilized to
distinguish four sounds during the heart cycle, namely,
S1, S2, S3, and S4, that are produced from particular
heart events such as the opening and closure of a valve.
8,9
The loudness of aforementioned sounds, however, varies
with the auscultation position.
10,11
Standard positions of
heart auscultation are the following: the mitral area
above the cardiac apex, the tricuspid area above the
fourth and fifth intercostal space along the left sternal
border, the pulmonic area above the second intercostal
space along the left sternal border and the aortic area
above the second intercostal space along the right sternal
border.
12
However, to assess heart function, the PCG should be
recorded and analysed correctly so abnormal heart
sounds related to potential cardiac abnormalities can be
detected. Namely, the PCG captured is always contami-
nated with noise and various artefacts. To make features
in the PCG more extractable, low-frequency noises
caused by respiratory movements and other environmen-
tal factors must be removed. Besides, high-frequency
noises such as heart murmurs must be filtered out. In this
regard, a PCG detection framework for heart sound anal-
ysis has been developed.
13
Using this detection frame-
work, the captured and analysed PCG included all four
heart sounds: S1, S2, S3 and S4. PCG can be described
by their intensity, pitch, position, quality and timing in
the heart cycle.
9
S1 comprises component M1 produced by the closure
of the mitral valve and component T1 produced by the
closure of the tricuspid valve. M1 is much louder than
T1 due to higher pressures in the left side of the heart.
M1 is heard at all standard auscultation positions (loud-
est at the apex), while T1 is usually only heard at the left
lower sternal border. When the heart rate is faster, S1 is
louder.
S2 comprises components A2, produced by the clo-
sure of the aortic valve, and P2, produced by the closure
of the pulmonic valve. A2 is much louder than the P2
due to higher pressures in the left side of the heart. A2 is
heard at all standard auscultation positions (loudest at the
right upper sternal border), while P2 is only heard at the
left upper sternal border.
S3 occurs in early diastole, just after S2, when the
mitral valve opens. S3 is produced during passive left
ventricular filling, when a large amount of blood strikes
a very compliant left ventricle. S3 may be sometimes
normal, but is often a sign of systolic congestive heart
failure. S3 is heard best at the apex. To hear S3, the bell
of the stethoscope is used.
S4 is produced when atrial contraction forces blood
through the atrioventricular valves striking the
noncompliant left ventricle. S4 occurs in late diastole
just before S1. S4 is almost always pathologic. It is often
a sign of diastolic heart failure or active ischemia. S4 is
best heard at the apex. These authors
13
have also identi-
fied the average heart sound durations, the time delay be-
tween them, and the duration of the heart cycles.
To enable an accurate heart sound signal analysis,
however, the normal frequency ranges of heart sounds
and murmurs were identified.
9,13
Namely, an accurate
distribution of frequencies in these sounds enables the
extraction of theoretical implications that might be use-
ful for clinical applications. In healthy subjects, the two
heart sounds, S1 and S2, are the most frequently used in
the PCG analysis. Namely, an absolute value of the la-
tency of S1 and S2 may help to determine the heart
sound type and detect abnormal heart sounds, while the
P. PE^LIN et al.: AUSCULTATION OF A HEART AND VASCULAR ACTIVITY DURING AURICULAR NERVE STIMULATION
92 Materiali in tehnologije / Materials and technology 57 (2023) 1, 91–100

amplitude of S1 may reveal information about myocar-
dial contraction ability.
14,15
Furthermore, the ratio be-
tween the S1 and the S2 amplitudes can provide an indi-
cation of heart function or malfunction. Finally, the ratio
of diastolic inter-phase (S2-S1) to systolic inter-phase
(S1-S2) can be used to detect insufficiency of a heart
blood supply. In practice, various transforms such as the
fast Fourier transform (FFT) are used for automatic heart
sound signal analysis.
16–18
In general, the human heartbeat is complex and
nonstationary artful body sound.
19
Its loudness is mea-
sured in decibels (dB) based on the current heart rate,
blood pressure, and distance to an examined person in a
completely silent room.
20
If the distance is to the person
is short, loudness might be within 5–10 dB. Korotkoff
sounds (KSs) are captured with a contact auscultation
transducer placed at the cubital fossa over the Brachial
artery simultaneously with the measurement of NIBP us-
ing the cuff and a sphygmomanometer.
21
When the cuff
is inflated to a pressure above the systolic Non Invasive
Blood Pressure (NIBP), the blood flow is completely oc-
cluded, thus no sound is audible. The sounds heard with
a stethoscope if placed distal to the blood pressure cuff
during the NIBP measurement are generated when the
cuff is gradually deflated and the blood flow through the
Brachial artery is rapidly enlarged. The KS, reported in
the literature,
22–25
have been classified into five phases:
1. The first sound heard as the cuff is released defines
the systolic pressure.
2. As the cuff is deflated, the blood flow in the artery is
increased and swishing sounds are produced.
3. The sounds heard become softer and louder or even
exceed the intensity of the sounds heard in Phase I.
4. As the cuff pressure is released further, the blood
flow in the artery becomes less turbulent and the
sounds heard are muffled and softer. This sound de-
fines the first reading of the diastolic pressure.
5. When the cuff pressure is released completely, KSs
disappear and the second reading of the diastolic
pressure is defined.
The frequency range of the KS reported in the litera-
ture has not been defined precisely. It was shown in ex-
periments, for instance,
26
that the KS’s energy was con-
centrated mainly below 400 Hz, while the passband of
the bandpass filter for the KS channel was 16 Hz to
400 Hz. It was shown in normal subjects, however, that
although most of the energy of the signal spectrum is be-
low 100 Hz,
27
a bandwidth of 20 Hz to 300 Hz is re-
quired for sufficient reproduction of the KS.
The motivation for the work was to develop and test a
system for the multichannel auscultation of different
types of biomedical signals capable of retaining the char-
acteristics of the main events within the signals, mini-
mizing the undesired noise and making the main events
more extractable. The first objective was to acquire and
compare the four PCGs that are produced during the
heart cycle from specific heart events such as the open-
ing and closing of valves with and without the tANS, re-
spectively. It was assumed that the system of four trans-
ducers developed was capable of differentiating types of
particular valve sounds that may be modified with the
tANS. The second objective was to acquire and compare
the KS signals generated by the blood flow in the
brachial artery during a standard NIBP measurement
with and without the tANS. It was assumed that KS
could provide some potentially useful information about
the state of the arteries before, during and after the
tANS. The main intention of the study was to develop
and experimentally test a system of acoustical transduc-
ers to be capable of capturing body sounds belonging to
heart and Korotkoff sounds for research purposes and not
to place it into service in a clinical environment. Due to
technical and less medical scope of the journal and the
background of the readers, the number of individuals
tested was limited to provide sufficient information on
P. PE^LIN et al.: AUSCULTATION OF A HEART AND VASCULAR ACTIVITY DURING AURICULAR NERVE STIMULATION
Materiali in tehnologije / Materials and technology 57 (2023) 1, 91–100 93
Figure 1: System of electronic stethoscope transducers

the technological and experimental procedures. As such
a statistical analysis could not be performed.
2 EXPERIMENTAL PART
2.1 Auscultation transducers
To build the multi-channel measuring set-up, various
custom-designed devices that provide the required mea-
suring techniques were used. Namely, to capture the
PCG that comes out from the heart valves in real-time,
four simple electronic transducers shown in Figure 1a
were developed. It should be noted that the approach
used for the development of the transducers was similar
to approaches already published by Hamid et al.,
6
and
Bhaskar.
28
However, the technological approaches and
materials used were different. For this purpose, basic sin-
gle metal-head acoustic stethoscopes having a diameter
of approximately 41 mm were instrumented with
low-noise microphone amplifier modules (GY-MAX
4466, MINGYUANDINGYE, China). The gain could be
adjusted between 25× and 125× using a small trimmer
potentiometer. This module comes with a wide band-
width omnidirectional electrostatic capacitor-based mi-
crophone (9767-52DB, TZT, Shenzhen, Beijing, China).
This module is best used for projects such as audio re-
cording/sampling and audio projects that use FFT. The
transducers developed can be used in the diaphragm or
the bell mode. The mode used can be selected by rota-
tion of the metal head by 180 degrees according to the
amplifier attached at the end of the tube, as shown in
Figure 1a. An optimum pressure was provided using
custom cut washers shown in Figure 1b that were ad-
hered between the stethoscopes and the skin. The trans-
ducers were connected to the data-acquisition system via
the connecting cable shown in Figure 1c to deliver an
analogue signal and to obtain the supply voltage.
The transducer for capturing the KS combined the
precision contact acoustical transducer (CM-01B,
Shenzhen Aimin Intelligent Technology Co., Ltd.,
Shenzhen, China) and signal amplifier module (CM-01,
Shenzhen Aimin Intelligent Technology Co., Ltd.,
Shenzhen, China). The CM-01B was a highly sensitive,
robust, lightweight and small-size contact acoustical
transducer based on a highly sensitive piezoelectric film
featuring an extremely low external noise interference
and wide frequency bandwidth. This combination was
attached to the custom-machined frame pieces that en-
abled additional suppression of the external noise inter-
ference. The finalized transducer for auscultation of the
KS is shown in Figure 2a. The transducer was connected
to the signal-amplifier module via the connecting cable
shown in Figure 2b. The combination delivers an ana-
logue signal to the data-acquisition system via the con-
necting cable shown in Figure 2c, while the cable that
provides DC voltage to the combination is shown in Fig-
ure 2d.
2.2 Calibration of the auscultation transducers
For the calibration of all the transducers developed, a
portable sound-pressure calibrator (Model: AWA6221A,
Hangzhou Aihua Instruments Co., Ltd., Zhejiang, China)
shown in Figure 3a was used. For the calibration, cou-
pling adapters that were compatible with each of the two
types of transducers were machined from plastics. Fig-
ure 3b shows a mechanical coupler for the PCG trans-
ducers, while Figure 3c shows a mechanical coupler for
the KS transducer. During calibration, each transducer
was adhered to the corresponding adapter using washers
so the transducer was faced directly to the portable
sound-pressure calibrator at a short distance. Finally, the
gain of the corresponding transducer was trimmed until
the same value as the normal sound pressure level of
P. PE^LIN et al.: AUSCULTATION OF A HEART AND VASCULAR ACTIVITY DURING AURICULAR NERVE STIMULATION
94 Materiali in tehnologije / Materials and technology 57 (2023) 1, 91–100
Figure 3: Setup for frequency-response testing
Figure 2: Transducer for the auscultation of KS

94 dB generated by the portable sound-pressure calibra-
tor was displayed. The transducers were calibrated under
normal room sound conditions and estimated to have a
peak-to-peak sound of approximately 15 dB.
2.3 Frequency-response tests of the auscultation trans-
ducers
To measure the inherent characteristics in each of the
four transducers developed, the frequency-response tests
within the range 20 Hz to 1000 Hz were conducted.
Through this testing, an answer to the question of
whether the frequency range covered by the transducer
was within the range that was appropriate for
auscultation of the PCG was obtained. During the re-
cording of PCG, the low-pass Butterworth filter that was
available for the high-performance, I/O data-acquisition
systems (DEWE-43, DEWESOFT d. o. o., Republic of
Slovenia) was selected. The purpose of the filtering was
to retain the information about the events within the cap-
tured signals, remove the undesired noise and make the
events more accessible. To accomplish the frequency-re-
sponse tests, the speaker shown in Figure 3d (SP-302
Mini Speaker for Notebook, Enzatec International Cor-
poration, Shenzhen, Guangdong, China), used as a sound
source, was mechanically coupled to each of the trans-
ducers.
To drive the speaker, a sine sweeping tone that
changed its frequency from 20 Hz to 1000 Hz was deliv-
ered from the Function/Arbitrary Waveform Generator
(Agilent 33120A, Santa Clara, California, U. S. A.). Af-
terwards, the tone was amplified using an audio ampli-
fier (LP-268B, Lepy, China, Lepy.com.cn) shown in Fig-
ure 3e. To accomplish the frequency-response testing, a
high-performance digital oscilloscope (Model: InstruStar
ISDS205A, Harbin ViMu Electronic Technology Co.,
Ltd., Harbin, China), was used. To accomplish a FFT,
however, the window-weighting function and the range
of the waveform data that provided the best results in
signal processing were selected. Afterwards, a cylindri-
cal adapter with a diameter of 60 mm and height of
10 mm made of machinable plastics, were adhered on
top of the speaker and each of the four transducers was
attached to the other side of an adapter using a washer.
2.4 Experimental procedures
To capture the PCG during the heart cycle in
real-time, four developed transducers were used. Pre-
cisely, to distinguish four types of sounds, i.e., S1, S2,
S3, and S4
10,12
an array of four transducers was placed at
the standard heart auscultation positions shown in Fig-
ure 4a. Accordingly, the first transducer was placed at
the aortic heart auscultation position located at the 2
nd
right intercostal space of the ribs near to the sternum.
The second transducer was placed at the pulmonic heart
auscultation position located at the 2
nd
left intercostal
space of the ribs close to the sternum. The third trans-
ducer was placed at the tricuspid heart auscultation posi-
tion located at the 4
th
left intercostal space of the ribs and
on the left lower sternal border. The fourth transducer
was placed at the mitral or apex heart auscultation posi-
tion located at the 5
th
left intercostal space of the ribs and
in the middle of the clavicular line. To reduce the effect
of high-frequency noises, the PCG signals were filtered
using a low-pass Butterworth filter, 3
rd
with a cut-off fre-
quency of 300 Hz. The system of four transducers was
used to differentiate the types of a particular valve
sounds that can be modified with the tANS.
The KSs were captured simultaneously with the mea-
surement of the NIBP using the second type of devel-
oped transducer. More descriptively, the transducer was
placed over the Brachial artery in the space between the
bottom of the cuff and the crease of the elbow, as shown
in the Figure 4b. It was presumed that Brachial arterial
sounds could provide some potentially useful informa-
tion about the state of the arteries before, during and af-
ter tANS. To compensate for the friction noise due to
body movements, washers were adhered to all the trans-
ducers.
For the tANS, the certified microprocessor-controlled
stimulator (Model SM9079, Shenzhen L-Domas Tech-
nology Ltd., Shenzhen, Guangdong, China) was used.
The pattern of the stimuli train was selected from the
palette of patterns available at the stimulator. The propri-
etary stimulating pulse selected was a symmetrical, cur-
rent-regulated rectangular and dual-directional stimulat-
ing pulse pair. The temporal parameters selected at the
touch screen of the stimulator were as follows:
– frequency: f = 3.3–45.5 Hz,
– stimulating phase duration: t
c
= 200 μs,
– anodic phase duration: t
a
= 200 μs,
– pulse train duration: 7.84 s,
– time gap between pulse trains: 1 s.
To deliver the stimuli to a particular position at the
external ear (EE), the negative output of the stimulator
P. PE^LIN et al.: AUSCULTATION OF A HEART AND VASCULAR ACTIVITY DURING AURICULAR NERVE STIMULATION
Materiali in tehnologije / Materials and technology 57 (2023) 1, 91–100 95
Figure 4: Positions of auscultation transducers

was connected to the corresponding electrode (cathode)
using the custom-developed switching unit. For the se-
lective tANS, the system shown in Figure 5 was devel-
oped. It was comprised of two silicone ear plugs, having
four platinum cathodes, each as shown in Figure 5b. The
plugs were then mounted onto the frame of the head-
phones, as shown in Figure 5a. During the trial, the
headphones were placed onto the head to fix the plugs
and to ensure as low as possible impedance between the
cathodes and the predefined stimulating positions at the
EE. The corresponding predefined positions of the cath-
odes, as shown in Figure 5c, are the following: pole, be-
low the pole, above the bottom and the bottom.
The reusable and self-adhering hydrogel common an-
ode (anode) (Model: 895220, Square (2" × 2"),
(50 × 50) mm), AXELGAARD MANUFACTURING
CO., LTD., Fallbrook, CA 92028, USA) was connected
to the positive output of the stimulator directly and
placed at the nape of the neck. To enable good electrical
and physical contact with the skin, all the stimulating po-
sitions and all the auscultation positions were degreased
with 70 % isopropyl alcohol and dried. During the tANS,
the intensity i
c
of the stimuli was set by the subject at a
level just below detected discomfort.
The trials were carried out under the same conditions
with a group of healthy female volunteers aged 23–25
years. All were free from any known cardiac disease. All
of them were in top physical and mental condition. We
carried out the study approved by the National Medical
Ethics Committee, Ministry of Health, Republic of
Slovenia, Unique Identifier No. 0120-297/2018/6).
The entire setup was connected to a volunteer, care-
fully considering the class safety standard (IEC 60601:
International Product Safety Standards for Medical De-
vices). All the components, including a PC, that were in
galvanic contact with the subject, were galvanically de-
coupled from the 220-V power line using a highest-qual-
ity, vintage, 500-VA isolation transformer (Mechanikai
Laboratórium (ML), Budapest, Hungary).
The conditions and exclusion criteria during the
tANS were as follows:
– The subject was instructed to avoid stress/tense and
physical activity before the trial.
– Smoking and/or drinking alcohol was prohibited be-
fore the trial.
– Consumption of a large meal was at the latest an
hour before the trial.
– The subject was asked to lay supine on the pedicure
chair with arms at about heart level to reduce hydro-
static pressure inaccuracies.
– The subject was asked not to talk and to remain as
still as possible to avoid motion artefacts.
A 65-minute trial was divided into four 10-minute se-
quences where i
c
was delivered separately onto each of
the four positions either at the left or the right EE, re-
spectively. The four sequences were separated with
5-minute shams where tANS was not delivered onto any
of the positions. The 65-minute trial also started and
ended with a 5-minute sham segment. Quantities i
c
, aor-
tic PCG (APCG), pulmonic PCG (PPCG), tricuspid PCG
(TPCG), mitral PCG (MPCG), and KS, were acquired
throughout the trial.
2.5 Signal acquisition and analysis
All six signals coming from the conditioning circuits
were gathered at 20 kHz with 24-bit resolution using the
above-mentioned data-acquisition system that had eight
differential analogue channel inputs and a USB 2.0 inter-
face. The stimulating intensity i
c
, however, was assessed
with a continuous measurement of the voltage drop
across the precision serial resistor of 10  connected to
the stimulator’s output and delivered to the data-acquisi-
tion system. During the acquisition, each signal was fil-
P. PE^LIN et al.: AUSCULTATION OF A HEART AND VASCULAR ACTIVITY DURING AURICULAR NERVE STIMULATION
96 Materiali in tehnologije / Materials and technology 57 (2023) 1, 91–100
Figure 5: System for multichannel tANS
Figure 6: Amplitude of frequency response determined by FFT

tered using the low-pass filter that was selected based on
the frequency band of the particular signal. Finally, data
were stored on a portable computer (Lenovo W541,
Lenovo, Beijing, China) to permit subsequent frequency
analysis, such as FFT, using software (DEWESoft 7.0.2).
3 RESULTS
Figure 6 shows the spectral amplitude of the fre-
quency response determined by FFT in one of the four
transducers developed for the auscultation of the PCG.
The figure shows that with a constant gain, the trans-
P. PE^LIN et al.: AUSCULTATION OF A HEART AND VASCULAR ACTIVITY DURING AURICULAR NERVE STIMULATION
Materiali in tehnologije / Materials and technology 57 (2023) 1, 91–100 97
Figure 8: Quantities recorded during selective tANS of pole position at the right EE
Figure 7: Quantities recorded without selective tANS

ducer developed was capable of providing a frequency
bandwidth wide enough to acquire the T1, T2, T3 and T4
into the PCG. More precisely, the transducer developed
was the most sensitive within the frequency range above
50 Hz and below 440 Hz.
The results also confirmed that the transducer devel-
oped based on a highly sensitive piezoelectric element,
remained the most sensitive within the frequency range
above 8 Hz and below 2.2 kHz, as declared in the list of
the product’s basic parameters. Figure 7 shows the re-
sults of experimental testing in which the tANS was not
applied. More descriptively, Figure 7 shows six frames
that visualize the quantities recorded: Figure 7a i
c
, Fig-
ure 7b APCG, Figure 7c PPCG, Figure 7d TPCG, Fig-
ure 7e MPCG, and Figure 7f KS where tANS was not
delivered to any position at the right EE.
Figure 8 shows the results of an experimental testing
in which tANS was delivered to the pole position on the
right EE. More descriptively, Figure 8 shows six frames
that visualize the quantities recorded: Figure 8a i
c
, Fig-
ure 8b APCG, Figure 8c PPCG, Figure 8d TPCG, Fig-
ure 8e MPCG, and Figure 8f KS when tANS was deliv-
ered onto the stimulating cathode at the pole position on
the right EE.
Table 1 shows an average value of heart cycle dura-
tions S1-S1 and S2-S2 as well as their amplitudes AS1
and AS2 without and during the tANS, respectively.
4 DISCUSSION
We presented the design, implementation and testing
of the system of transducers for multichannel auscul-
tation of a heart and vascular activity dynamics during
the tANS of the EE. Besides a precise description of the
system developed and techniques used for the capture
and recording, we also review the preliminary results of
a PCG and KS analysis with regard to the potential ef-
fects of the tANS. The motivation for the work was the
desire to develop and test the system capable of retaining
the characteristics of the main events within the PCG and
KS, to minimize the undesired noise and make the main
events effectively extractable. Namely, modern biomedi-
cal signal-processing techniques are able to accurately
characterize significant features of the heart sounds con-
tained within the PCGs
29,30
as well as features of the
KS.
31,32
However, in most of the devices developed, the
loudness of captured PCG and KS, however, varied sig-
nificantly with the auscultation positions.
11
Since both
the PCG and KS were in a lower part of the frequency
range, the frequency response of the transducer devel-
oped should enable an accurate capture within an as-
signed frequency region.
The results of testing in the human model showed
that both transducer types developed were appropriate
for capturing the PCG and KS, respectively. The results
showed that all of the four transducers developed as the
PCG are optimized for the auscultation of body sounds
coming from heart activity. The results also showed that
the transducer developed based on a highly sensitive pi-
ezoelectric element was appropriate for the auscultation
of KS coming from the Brachial artery. The objective to
acquire and compare the four PCGs that are produced
during the heart cycle from particular heart events with
and without the tANS was achieved. Similarly, the objec-
tive to acquire and compare the KS signals generated by
the blood flow in the Brachial artery during a standard
NIBP measurement with and without the tANS was
achieved. However, the assumption that transducers de-
veloped for the auscultation of PCG were capable of dif-
ferentiating types of particular valve sounds potentially
modified with the tANS was barely confirmed.
Table 1 shows that both durations S1-S1 and S2-S2
were slightly shortened during the tANS. It means that
during the tANS, heart cycles became slightly shorter
and thus, the heart rhythm slightly higher. It was noticed,
however, that in both TPCG and MPCG the peak values
of S1 are larger than the peak values of S2 without as
well as during the tANS. It was also noticed that in both
APCG and PPCG the peak values of AS2 are larger than
the peak values of AS1 without as well as during the
tANS. In general, both amplitudes of AS1 and AS2 be-
came slightly higher during the tANS. It means that dur-
ing the tANS both heart sounds S1 and S2 became
louder. The assumption that the transducer developed for
the auscultation of KS was capable of differentiating
types of sounds potentially modified with the tANS was
barely confirmed. Similarly, the assumption that KS
could provide some potentially useful information about
the state of the arteries before, during and after tANS
was barely confirmed.
Compared to transducers developed by other re-
searchers, the transducers developed within the study
provided an optimum pressure against the skin. This was
obtained using custom cut washers that were adhered to
the additional aluminium ring mounted at the stetho-
scope head and the skin at the same time. As the washers
fixed the transducers onto the skin more firmly, the gen-
P. PE^LIN et al.: AUSCULTATION OF A HEART AND VASCULAR ACTIVITY DURING AURICULAR NERVE STIMULATION
98 Materiali in tehnologije / Materials and technology 57 (2023) 1, 91–100
Table 1: Average value of durations S1-S1, S2-S2 and amplitudes AS1 and AS2 without and during tANS
WITHOUT tANS DURING tANS
S1-S1 (s) S2-S2 (s) AS1 (dB) AS2 (dB) S1-S1 (s) S2-S2 (s) AS1 (dB) AS2 (dB)
APCG NA 1.115 NA 32.067 NA 0.987 NA 36.062
PPCG NA 1.112 NA 32.45 NA 0.987 NA 34.7
TPCG 1.115 NA 26.205 NA 1.026 NA 32.475 NA
MPCG 1.122 NA 19.74 NA 0.968 NA 18.666 NA

eration of friction noise during body movements and the
capture of external sounds and noises were minimized.
3,4
As a result, captured body sounds were made more inter-
pretable in an offline review and analysis. However, the
lower and upper limits of the frequency bandwidth be-
tween 50 Hz and 440 Hz depended mainly on the vibra-
tional performance of membrane at the stethoscope.
Good results regarding the suppression of external
and friction noise were also obtained with the transducer
developed for the auscultation of KS coming from the
Brachial artery.
All the conditions were kept as steady as possible in
all the trials. However, the trials were carried out and
analysed by one investigator, which might have led to
some biases. In addition, the numerical data extracted
from the captured signals could enable a subjective inter-
pretation.
The directions of our future work will include a pre-
cise identification of the durations of S1, S2, S3 and S4,
the delay S3-S2, the delay S4-S1, the duration of heart
cycles, the systolic period and the diastolic period.
13
Our
future work will also include a precise identification of
the normal frequency ranges of S1, S2, S3, S4 and the
frequency range of murmurs.
9
However, for a more de-
tailed analysis of potential tANS effects on the quantities
extracted from the PCG and KS, computational models
will be necessary.
Finally, the directions of our future work will include
tANS at even more sites on the EE and fine-tuning of the
stimulation parameters to improve the stimulating elec-
trode-skin contact.
5 CONCLUSIONS
In regard to the auscultation of PCG, the results con-
firmed that the system of four transducers was capable of
differentiating types of particular valve sounds and that
they can be modified with the tANS. It could detect the
accurate positioning of S1 and S2 and potentially enable
the extraction of the information hidden in the S1 and S2
spectra that are beneficial for the evaluation of clinical
heart function.
In regard to the auscultation of KS, however, the re-
sults confirmed that Brachial arterial sounds could pro-
vide some potentially useful information about the state
of the arteries before, during and after tANS.
However, the number of the subjects tested and the
measurements performed were low. In this relation, re-
sults representing changes in PCG and KS should not be
considered as a basis to derive credible conclusions, but
should be considered as a basis for further research activ-
ities.
The presented work has implications for the setup de-
sign and assessment of the efficiency of multichannel
tANS to modify heart-activity dynamics.
The system for multichannel auscultation of heart
function dynamics developed can assist cardiologists in
taking appropriate actions to diagnose the initial stage of
cardiovascular disorders. The system can also be used as
a support for future research in various pathological con-
ditions.
The system was developed primarily for research and
not for commercial purposes. To place a system into ser-
vice in clinical environment, however, the system should
be tested in a larger group of subjects and in more trials.
Once testing was completed, the system could obtain the
CE Marking, ensuring that it meets all the requirements
of the Medical Device Regulation (EU) 2017/745.
Acknowledgement
This research was funded by the Slovenian Research
Agency, Ministry of Education, Science and Sport,
Ljubljana, Republic of Slovenia, Grant Number P3-0171,
which was awarded to the Institute of Pathophysiology,
Faculty of Medicine, University of Ljubljana, Republic
of Slovenia.
6 REFERENCES
1
M. Elgendi, TERMA Framework for Biomedical Signal Analysis:
An Economic-Inspired Approach, Biosensors, 55 (2016)6 ,
doi:10.3390/bios6040055
2
T. Rahman, A. T. Adams, M. Zhang, E. Cherry, B. Zhou, H. Peng, T.
Choudhury, BodyBeat: a mobile system for sensing nonspeech body
sounds, Proc. of the 12
th
annual international conference on Mobile
systems, applications, and services, Bretton Woods 2014, 2–13,
doi:10.1145/2594368.2594386
3
S. Nirjon, R. F. Dickerson, P. Asare, Q. Li, D. Hong, J. Stankovic, P.
Hu, G. Shen, X. Jiang, Auditeur: A mobile-cloud service platform
for acoustic event detection on smartphones, MobiSys., 13 (2013),
403–416
4
K. Yatani, K. Truong, Bodyscope: A wearable acoustic sensor for ac-
tivity recognition, Proc. of the 2012 ACM Conference on Ubiquitous
Computing, Pittsburgh 2012, 341–350, doi:10.1145/2370216.
2370269
5
M. Elgendi, R. Fletcher, Y. Liang, N. Howard, N. H. Lovell, D.
Abbott, K. Lim, R. Ward, The use of photoplethysmography for as-
sessing hypertension, npj Digit. Med., 2 (2019) 60, doi:10.1038/
s41746-019-0136-7
6
M. A. A. Hamid, M. Abdullah, N. A. Khan, Y. M. A. AL-Zoom,
Biotechnical System for Recording Phonocardiography, International
Journal of Advanced Computer Science and Applications (IJACSA),
10 (2019) 8, doi:10.14569/IJACSA.2019.0100864
7
P. S. Vikhe, N. S. Nehe, V. R. Thool, Heart Sound Abnormality De-
tection Using Short Time Fourier Transform and Continuous Wavelet
Transform, Proc. of the 2009 Second International Conference on
Emerging Trends in Engineering & Technology, Nagpur 2009,
50–54, doi:10.1109/ICETET.2009.112
8
A. Ramovi}, L. Bandi}, J. Kevri}, E. Germovi}, A. Subasi, Wavelet
and Teager energy operator (TEO) for heart sound processing and
identification, Proc. of the CMBEBIH 2017, Sarajevo 2017, 495–502
9
https://www.healio.com/cardiology/learn-the-heart/cardiology-re-
view/topic-reviews/heart-sounds
10
R. J. Oweis, H. Hamad, M. Shammout, Heart Sounds Segmentation
Utilizing Teager Energy Operator, Journal of Medical Imaging and
Health Informatics, 4 (2014) 4, 488–499, doi:10.1166/jmihi.
2014.1292
P. PE^LIN et al.: AUSCULTATION OF A HEART AND VASCULAR ACTIVITY DURING AURICULAR NERVE STIMULATION
Materiali in tehnologije / Materials and technology 57 (2023) 1, 91–100 99

11
M. McGregor, M. B. Rappaport, H. B. Sprague, A. L. Friedlich, The
calibration of heart sound intensity, Circulation, 13 (1956)2 ,
252–256, doi:10.1161/01.cir.13.2.252
12
C. Ahlström, Processing of the Phonocardiographic Signal-Methods
for the Intelligent Stethoscope, Linköping Studies in Science and
Technology, Thesis No. 1253, Linköping University, Institute of
Technology, Linköping 2006
13
V. N. Varghees, K. I. Ramachandran, A novel heart sound activity
detection framework for automated heart sound analysis, Biomedical
Signal Processing and Control, 13 (2014), 174–188, doi:10.1016/
j.bspc.2014.05.002
14
T. H. Chowdhury, K. N. Poudel, Y. Hu, Time-Frequency Analysis,
Denoising, Compression, Segmentation, and Classification of PCG
Signals, IEEE Access, 8 (2020) 160882–160890, doi:10.1109/ac-
cess.2020.3020806
15
J. Zhong, F. Scalzo, Automatic heart sound signal analysis with re-
used multi-scale wavelet transform, International Journal of Engi-
neering Science, 2 (2013) 50, 50–57
16
B. M. Whitaker, P. B. Suresha, C. Liu, G. D. Clifford, D. V. Ander-
son, Combining sparse coding and time-domain features for heart
sound classification, Physiol. Meas., 38 (2017) 8, 1701–1713,
doi:10.1088/1361-6579/aa7623
17
P. Langley, A. Murray, Heart sound classification from unsegmented
phonocardiograms, Physiol. Meas., 38 (2017) 8, 1658–1670,
doi:10.1088/1361-6579/aa724c
18
R. N. Bracewell, The Fourier transform, Scientific American, 260
(1989) 6, 86-89, 92-95, doi:10.1038/scientificamerican0689-86
19
J. Zhong, F. Scalzo, Automatic heart sound signal analysis with re-
used multi-scale wavelet transform, International Journal of Engi-
neering and Science, 2 (2013) 7, 50–57
20
P. Malema, Watch: the world’s quietest room is so quiet, you can
hear the sound of your heart beat, https://www.jacarandafm.com/
shows/workzone-with-barney-simon/watch-worlds-quiet-
est-room-so-quiet-you-can-hear-sound-your-heart-beat/
21
E. P. McCutxheon, R. F. Rushmer, Korotkoff Sounds an experimental
critique, Circulation Research, 20 (1967) 2, 149–161, doi:10.1161/
01.res.20.2.149
22
https://trovalarisposta.com/biblioteca/articolo/read/107225-what-are-
korotkoff-sounds-and-what-do-they-represent
23
https://en.wikipedia.org/w/index.php?title=Korotkoff_sounds&
oldid=1101143114
24
M. Al-Amri, M. Nizar, A. Alzaben, M. Fadlee, Characteristics of
Korotkoff sounds using instantaneous frequency signal analysis,
WSEAS Transactions on Signal Processing, 3 (2007), 296–302,
https://orca.cardiff.ac.uk/id/eprint/43767
25
B. Wilk, R. Hanus, Blood flow in the brachial artery compressed by
a cuff, EPJ Web of Conferences, 213 (2019), 02099, doi:10.1051/
epjconf/201921302099
26
H. Xiang, Y. Liu, Y. Qin, Z. Cao, T. Guo, M. Yu, A pilot application
of Korotkoff sound delay time in evaluating cardiovascular status,
Technology and Health Care, 23 (2015), S419–S426, doi:10.3233/
THC-150978
27
C. D. Kapse, B. R. Patil, Auscultatory and Oscillometric methods of
Blood pressure measurement: a Survey, International Journal of En-
gineering Research and Applications (IJERA), 3 (2013) 2, 528–533
28
A. Bhaskar, A simple electronic stethoscope for recording and play-
back of heart sounds, Adv. Physiol. Educ., 36 (2012), 360–362,
doi:10.1152/advan.00073.2012
29
A. M. Noor, M. F. Shadi, The heart auscultation. From sound to
graphical, Journal of Engineering and Technology (JET), 4 (2013)2,
73-84
30
C. N. Gupta, R. Palaniappan, S. Rajan, S. Swaminathan, S. M.
Krishnan, Segmentation and classification of heart sounds, Proc. of
the Can. Conf. Electrical Comput. Eng., Saskatoon 2005,
1674–1677, doi:10.1109/CCECE.2005.1557305
31
D. A. Sykes, K. McCarty, E. Mulkerrin, D. J. Fisher, J. P. Woodcock,
Correlation between Korotkoff’s sounds and ultrasonics of the
brachial artery in healthy and normotensive subjects, Clin. Phys.
Physiol. Meas., 12 (1991) 4, 327–331, doi:10.1088/0143-0815/
12/4/002
32
W. C. Kao, C. C. Wei, Automatic phonocardiograph signal analysis
for detecting heart valve disorders, Expert Systems with Applica-
tions, 38 (2011) 6, 6458–6468, doi:10.1016/j.eswa.2010.11.100
P. PE^LIN et al.: AUSCULTATION OF A HEART AND VASCULAR ACTIVITY DURING AURICULAR NERVE STIMULATION
100 Materiali in tehnologije / Materials and technology 57 (2023) 1, 91–100