ArticlePDF Available

Vibrocarotidography: A Novel Measurement Technique to Quantify Pulsations at Common Carotid Arteries

Authors:

Abstract and Figures

In this paper, aortic ejected blood-flow and aortic pressure is investigated as an independent tool for diagnosing cardiovascular risk. This study presents vibrocarotidography (ViCG), a novel noninvasive and nonintrusive way to measure aortic blood-flow variations in each heartbeat through carotid arteries. Most of the existing state-of-the-art works suggested to use contact-based pressure sensors and non-contact sensing devices, including wave radar and laser Doppler vibrometer (LDV) for carotid pulse acquisition. However, these sensors have operational design limitations, and poor immunity against environmental noises. To address these issues, the proposed method uses a miniaturized and cost-efficient micro-electro-mechanical system (MEMS)-based accelerometer sensor to record the vibrational pulsations on common carotid artery. The paper presents our developed electronic circuitry for ViCG signal acquisition and signal processing perspective for estimating indispensable cardiac events. The study focuses to show the ViCG signal as an alternative measure of central blood flow variations. Significance of the ViCG signal is exhibited by assessing the rate of pulsations and comparing with the heart-rhythms measured from the reference ECG and PPG signals. A quantitative Bland-Altman analysis shows a mean difference of-0.01 ms and correlation coefficient of 0.93 (R-squared) between the cardiac intervals measured from the ViCG and ECG signals. Whereas, they are found to be 0.03 ms and 0.92 for the ViCG-PPG signal pair. They reveal a highly strong correlation and agreement for heart cycle estimation. The performance analysis suggest that the ViCG signal acquired through a simple MEMS-based accelerometer can be utilized as a surrogate of central blood flow measurement and may be employed for continuous health monitoring in personalized-, home-, and hospital-healthcare systems.
Content may be subject to copyright.
0018-9456 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIM.2021.3115203, IEEE
Transactions on Instrumentation and Measurement
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2021 1
Vibrocarotidography: A Novel Measurement
Technique to Quantify Pulsations at Common
Carotid Arteries
Tilendra Choudhary, Member, IEEE, Mousumi Das, M.K. Bhuyan, Senior Member, IEEE, and L.N. Sharma
Abstract—In this paper, aortic ejected blood-flow and aortic
pressure is investigated as an independent tool for diagnosing
cardiovascular risk. This study presents vibrocarotidography
(ViCG), a novel noninvasive and nonintrusive way to measure
aortic blood-flow variations in each heartbeat through carotid ar-
teries. Most of the existing state-of-the-art works suggested to use
contact-based pressure sensors and non-contact sensing devices,
including wave radar and laser Doppler vibrometer (LDV) for
carotid pulse acquisition. However, these sensors have operational
design limitations, and poor immunity against environmental
noises. To address these issues, the proposed method uses a
miniaturized and cost-efficient micro-electro-mechanical system
(MEMS)-based accelerometer sensor to record the vibrational
pulsations on common carotid artery. The paper presents our
developed electronic circuitry for ViCG signal acquisition and
signal processing perspective for estimating indispensable cardiac
events. The study focuses to show the ViCG signal as an alterna-
tive measure of central blood flow variations. Significance of the
ViCG signal is exhibited by assessing the rate of pulsations and
comparing with the heart-rhythms measured from the reference
ECG and PPG signals. A quantitative Bland-Altman analysis
shows a mean difference of -0.01 ms and correlation coefficient
of 0.93 (R-squared) between the cardiac intervals measured from
the ViCG and ECG signals. Whereas, they are found to be 0.03
ms and 0.92 for the ViCG-PPG signal pair. They reveal a highly
strong correlation and agreement for heart cycle estimation.
The performance analysis suggest that the ViCG signal acquired
through a simple MEMS-based accelerometer can be utilized
as a surrogate of central blood flow measurement and may
be employed for continuous health monitoring in personalized-,
home-, and hospital-healthcare systems.
Index Terms—Vibrocarotidography (ViCG), Micro-electro-
mechanical system (MEMS), Cardiac cycle interval.
I. INTROD UC TI ON
By evolving day-by-day, noninvasive and nonintrusive
means of physiological sensing for the cardiac assessment
has become ubiquitously popular and a vital need for both
personalized and hospital healthcare systems. It helps to im-
plement effective and continuous monitoring of cardiac events
for the early detection of cardiovascular diseases (CVDs).
Various cardiac activities that can be considered essential
biomarkers for predicting cardiovascular disorders are valves’
opening and closing instants, blood ejection and filling in heart
chambers, pumping force, and blood flow variations at central
and peripheral arteries. Blood pulsations caused due to aortic
T. Choudhary, M. Das, M.K. Bhuyan, and L.N. Sharma are with the
Department of Electronics and Electrical Engineering, Indian Institute of
Technology Guwahati, India-781039 (e-mails: {tilendra, mousumi18a, mkb,
lns}@iitg.ac.in).
pressure at different body locations can be measured by vari-
ous sensing devices. Usually, pressure sensors are used for the
assessment of peripheral blood-flow on the radial or brachial
artery. Clinically, this assessment is accomplished through
sphygmomanometers and oscillometric devices. However, due
to peripheral amplification, the pressure induced by blood flow
at peripheral arteries is always different from that of the origin
artery, i.e., aorta and therefore, it is suggested to assess the
central (or aortic) blood-flow instead of peripheral ones [2]–
[6], [17]. Especially in the cases of antihypertensive medica-
tions consumption, the peripheral blood-flow information is
not suggested to predict the cardiovascular-events. Thus, the
central blood-flow measurement is highly recommended for
the assessment of aortic health, heart pumping force, blood
turbulence level, valvular functions, and many other cardiac
health related issues. Among all the measurement sites, the
skin overlying the common-carotid near thyroid cartilage is
found suitable to obtain the cardiac pulsations [7]. The pulsa-
tions at the carotid artery are caused due to the variations in
the aortic blood flow. Since the carotid artery, which is a major
branch of aorta, has orifice of relatively small diameter and it
resides near to the skin surface, high pulsations are perceived
on the site. This phenomenon shows the appropriateness of
carotid artery for measuring central blood flow variations. One
of the oldest technique used for the measurement of carotid
pulse information is catheterization [8]. But due to its invasive
nature, it may lead to some clinical problems, such as bleeding,
infection, and ischemia, thus causing inconvenience to the
users. Additionally, it requires a standard clinical setup and an
expert. For these reasons, new techniques have been evolved to
measure carotid pulse information in a more comfortable way.
With the advancement in technology, various sensing devices
have been developed to noninvasively capture the pulsations
generated at carotid arteries.
Among all the sensing modalities, pressure transducers
were widely used to acquire the carotid pulses noninvasively
[9]–[13]. In this category, carotid pulse pressure waves are
recorded noninvasively by the use of applanation tonometry
[11]. It is performed by a little flattening of the carotid
artery with the help of a pressure sensor. Acquiring a reliable
pulsatile signal using this method is a challenging task as it re-
quires precision in the compression of the artery by the sensor.
Additionally, the pressure sensors are not suitable for long-
term monitoring of carotid signal because compressing the
measurement site, i.e., neck surely causes inconvenience to the
user. Another technique has exploited radio-frequency based
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY GUWAHATI. Downloaded on September 24,2021 at 07:09:25 UTC from IEEE Xplore. Restrictions apply.
0018-9456 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIM.2021.3115203, IEEE
Transactions on Instrumentation and Measurement
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2021 2
Fig. 1. Frontal anatomical view of human torso including heart, rib cage,
thyroid cartilage, common-carotid artery, left and right sided internal and
external carotid and their bifurcation points, and sternum. The ViCG sensor
location point is also highlighted on the neck surface.
ultrasound (US) imaging of the heart to estimate the carotid
pulses [14]–[18]. Initially, the carotid diameter waveform is
obtained through the US image, and then, calibrated with the
pressure. However, the technique is limited due to the fact that
it relies on a heavy and bulky equipment, which need skilled
medical experts. Recently, Wang et al. [19] introduced confor-
mal and stretchable ultrasonic devices that facilitate continuous
and noninvasive monitoring of cardiac pulsations below the
human skin. The method uses multiple piezoelectric ultrasound
transducers (PUTs) in matrix-array format and a total of 90
signals are used to reconstruct the image. However, due to
its ultrathin and stretchable design, the developed wearable
patch is comfortable to use. In the literature, various non-
contact carotid signal recording devices were also proposed,
such as continuous wave radar and laser Doppler vibrometer
(LDV) [20]–[25]. However, these techniques are susceptible
to noises and need more precision in signal acquisition steps.
To overcome the limitations of the state-of-the-art approaches,
we introduced a new way to measure the aortic mechanical
vibrations from the common carotid using a micro-electro-
mechanical system (MEMS) based accelerometer in our ear-
lier filed patent [26]. The fundamental principle behind this
measurement is very much similar to the signal acquisition
of seismocardiogram (SCG) signal, which represents cardiac
induced chest wall vibrations [26], [27]. The MEMS-based
accelerometer sensor can efficiently detect the vibrations on
the carotid arteries and it may prove to be a robust sensing
device among all the other modalities. With the use of a
MEMS-based accelerometer, our aim is to capture the aortic
micro-vibrations caused due to rapid blood ejection phase
in each heart cycle. This is implicitly achieved by sensing
the internal vibrations perceived at the carotid skin-surface.
The signal acquired using this novel noninvasive approach is
termed as vibrocarotidogram (ViCG).
This proposed signal acquisition method provides a great
opportunity to study the central blood flow phenomenon in
a simple and noninvasive manner. To be more specific, our
designed and developed electronic circuitry and its operational
functionality for signal acquisition have been encapsulated
in [26]. Whereas, this paper mainly focuses towards signal
processing aspects, analysis and applications for establishing
Fig. 2. Simultaneously recorded ECG, ViCG, and PPG cycles. (a) ECG,
(b) dorso-ventral ViCG signal, and (c) PPG signal. (Abbreviations used– VP:
ViCG pulse peak, BTI: Blood turbulence interval, PTT: Pulse transit time, F:
Foot, SP: Systolic peak, DN: Dicrotic notch, and DP: Diastolic peak.)
the emerging ViCG signal. The rest of the paper is organized
as follows: in Section II, we present the physiological back-
ground and characteristics of ViCG signal. Section III presents
the utilized materials and our methods. In Section IV, the
experimental results and performance evaluation is presented.
Finally, conclusions are drawn in Section V.
II. PHYSIOLOGY AND SIGNA L CHA RACTERISTICS
As a key organ of blood circulatory system, the heart
supplies oxygenated blood and other important nutrients to all
body parts for their proper functioning. It is comprised of four
chambers, left and right, atria and ventricles. The oxygenated
blood from the left ventricle of the heart is pumped out from
the largest artery, i.e., aorta, and is carried to all the body
parts through branches of aorta called arteries. Aortic ejected
blood-flow through the arteries results in rhythmic pulsations
that can be sensed on the peripheral skin surface. The major
branch of the aorta, i.e., common carotid show significantly
high pulsations among all the arteries. The observed pulsations
at the carotid corresponds to the central blood flow. The
carotid arteries are basically supported by sternocleidomastoid
muscles, whereas, the blood circulation on these arteries is
regulated by jugular veins [28]. It is originated from the
aortic arch and is situated on both the left and right sides
of the thyroid cartilage as shown in Fig. 1. The oxygenated
blood is supplied through these arteries to the facial tissues
and brain regions. These arteries are bifurcated into internal
and external carotid arteries at the level of top-border of
the thyroid cartilage. A dilated area at the bifurcation of
the common carotid produces significant pulsations due to
rhythmic variation of blood supply from the aorta. Therefore,
this specific location is found suitable for sensor placement.
Fig. 1 shows the frontal anatomical view of human torso
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY GUWAHATI. Downloaded on September 24,2021 at 07:09:25 UTC from IEEE Xplore. Restrictions apply.
0018-9456 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIM.2021.3115203, IEEE
Transactions on Instrumentation and Measurement
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2021 3
TABLE I
CHARACT ER IZATION O F DIFF ER ENT C AR DIAC SIG NAL S
Signal Phenomenon captured Sensing device Represent activity
ECG Electrical Ag-AgCl Electrodes Heart electrical conduction
PPG Mechanical Optoelectric finger-cuff Peripheral blood flow
ViCG Mechanical Accelerometer Central blood flow
along with sensor location to capture resultant vibrations due
to aortic blood flow. Since the vibration signal is generated
from the carotid arteries, it is named as vibrocarotidogram
(ViCG). Fig. 2 shows the annotated ViCG signal along with the
simultaneously recorded electrocardiogram (ECG) and photo-
plethysmogram (PPG) signals. These signals are also described
in terms of their nature of physical phenomena captured,
sensors used, and their cardiac tracking abilities in Table I.
The ViCG signal is a series of periodic pulsatile structures
emerging in each heart cycle, which implicitly conveys the
information of aortic blood-flow variations. As the ViCG is
acquired just near to the heart, it contains rich information
related to central blood flow and heart activities, such as
rhythmic nature, region and level of aortic blood turbulence,
and valvular actions in each cardiac cycle. This makes it
different from the existing ECG and PPG modalities. The PPG
signal may also be used to approximate the central blood flow
information; however, this information gets deteriorated due
to far peripheral-site signal acquisition, technical problems,
and other anatomical factors [19]. Also, PPG has limited
penetration depth (<8 mm), which makes it unsuitable for
central blood flow measurement at carotid that is embedded in
a depth of around 3 cm. Additionally, both the signals differ
by their sensing mechanisms. That is why, although the ViCG
and the PPG signals are the outcome of the same physiological
process, they are not morphologically similar. The peaks of the
ViCG pulsatile waves correspond to the rhythmic pulsations
on the common carotid and are annotated as VPs (peak of
ViCG wave). The generation of VP point is mainly caused
due to maximum blood turbulence in each heartbeat. Hence,
the maximum blood turbulence phenomenon of the aorta can
be characterized by the pulsatile waves in the ViCG signal.
The characteristics of pulsatile waves, including VP peaks
and time intervals like VT1-to-VT2and VP-to-VP can be used
to assess the performance of the carotid, aorta, and heart.
Although PPG reflects the blood flow information at peripheral
sites, it is unable to provide true central blood flow information
due to relying on many constraints like artery length, diameter,
stiffness, effective cross-section area, and blood viscosity. True
modeling of aortic blood flow for these dynamic parameters
is another level challenge. As a solution, the ViCG is able
to provide the central blood flow information with ease. The
amplitude of VP in ViCG measures the strength of blood
turbulence, while its time-instant indicates its rhythmic oc-
currence. Pulse rate (PR) information could also be measured
from the consecutive VP points. It closely resembles to heart
rate information. The PR is expressed as:
P Rk= 60/Tk(1)
where, k= 1,2,· · · , K, and Tdenotes cardiac interval that
is calculated as the difference between the consecutive VP
Fig. 3. Spectro-temporal analysis of ViCG signal (a) ViCG, (b) Magnitude
spectrum, (c), (d) and (e) Wideband, midband and narrowband spectrograms,
respectively.
time instants, i.e.,
Tk=V PkV Pk1,k= 1,2, ...K (2)
where, V Pkdenotes the time instant of kth VP point. The
blood turbulence increases dramatically as soon as the aortic
valve opens, which in turn generates vibrations on the carotid
arteries. The aortic valve opening is the mechanical effect
of ventricular depolarization of the heart during electrical
conduction. That electrical phenomenon is mapped with R-
wave of the ECG signal. Hence, the time-interval between
R-peak in the ECG and VP-peak in the ViCG can be used to
estimate the time taken for the pulsations to travel from the
heart to the carotid. This is termed as blood turbulence interval
(BTI) and it can be expressed as:
BT Ik=V PkRk(3)
Thus, these parameters obtained from the ViCG signal can
be used for cardiac health assessment. Along with temporal
information, the signal can also be characterized in spectral
domain. Fig. 3(a) and (b) show the ViCG signal and its
corresponding magnitude spectrum, respectively. Whereas, (c),
(d), and (e) presents its time-frequency analysis via wide-,
mid- and narrow-band spectrograms with window size of 64,
128, and 256 ms, respectively. With vertical striations, wide-
band spectrogram indicates cardiac periodicity exhibited by
the ViCG signal. While the fine distinctions of ViCG spectral
components can be made from the narrowband spectroscopic
ViCG. The significant spectral components reside within 10
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY GUWAHATI. Downloaded on September 24,2021 at 07:09:25 UTC from IEEE Xplore. Restrictions apply.
0018-9456 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIM.2021.3115203, IEEE
Transactions on Instrumentation and Measurement
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2021 4
Fig. 4. Schematic diagram showing pre-processing steps for the recorded ViCG signal
Fig. 5. Experimental setup for recording of signals. (1) ViCG sensor, (2) PPG
sensor, (3) ECG electrodes, (4) ViCG conditioning circuitry, (5) Biopac DAQ
system, and (6) display device.
Hz frequency. Thus, the ViCG signal can be categorized as a
low-frequency signal, and a suitable band can be selected for
time-frequency representation of this signal as per the required
applications.
III. MATERIALS AND METHODS
A. ViCG sensor
In the proposed work, turbulence of the blood on the
common-carotid is measured in a new way by using a
MEMs-based tri-axial accelerometer sensor. It captures the
accelerations caused by shocks, motions, or vibrations. How-
ever, only the z-axis measurement i.e., orthogonal direction
to the measurement site has been considered for the study
because significant variations are observed in the z-direction
as compared to x- and y-directions. To record the ViCG
signal, the sensor is mounted on the neck surface slightly
left/right side of the thyroid cartilage at the location of the
carotid bifurcation. The vibrations induced in each heartbeat
change the differential capacitances of the sensor due to its
electro-mechanical mechanism, which results in sensor voltage
outputs. The output voltage value is directly proportional to
the vibrations produced due to aortic ejected blood.
B. Experimental set-up and processing
For our experimentation purpose, the ViCG along with
concurrent ECG and PPG signals were recorded for two
minutes from each of the twenty healthy subjects (sixteen
males and four females). Recording process was performed
Fig. 6. Block diagram of the proposed ViCG peak detection algorithm.
at the EMST lab of IIT Guwahati with proper consent of
volunteers. For all the subjects, recordings were done for
ECG in Lead-II configuration using Ag-AgCl electrodes, z-
axis channel of ViCG using our self built signal acquisition
electronic device (refer [26]), and PPG at fingertip using Nonin
medical’s SenSmart®optoelectronic finger clip sensor. All the
signals are sampled at 1 kHz frequency and synchronized
using MP150 DAQ system (BIOPAC Systems, Inc.).
To eliminate the environmental interferences from the
recorded ViCG signal, the measured data is pre-processed in
a sequential order using an electronic circuitry. The process of
acquiring ViCG using our conditioning circuitry is schemati-
cally shown in Fig. 4 and the experimental setup for signals’
recording is shown in Fig. 5. The electronic circuitry is used
for enhancement, selection of bandwidth, and analog to digital
conversion of the acquired ViCG signal. The recorded ViCG
signals are electronically pre-amplified to enhance the signal
for further analysis. The amplified signal are then filtered using
an active low pass filter with a cut-off frequency of 10 Hz. A
capacitor-based DC blocker is used to remove the DC offsets
from the filtered signals. The DC blocker is further interfaced
with an analog-to-digital converter via a buffer circuit to
digitize the analog signals. A good quality ViCG signal is
achieved by appropriate selection of the sampling rate and the
number of quantization bits.
Once the signals are achieved from the main processing
unit, the VP peaks are detected on the signals to analyse
the cardiac cycles. Fig. 6 shows the block diagram of the
proposed ViCG peak detection technique. Initially, mean is
removed and amplitude is normalized for the acquired digital
ViCG signal. Then, a high pass filter (HPF) with a cut-off
frequency of 1 Hz is used to remove the baseline drift from
the signal. Blood turbulence region of the filtered ViCG signal
is emphasized using an envelope construction scheme. The
desired task is accomplished by constructing an absolute upper
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY GUWAHATI. Downloaded on September 24,2021 at 07:09:25 UTC from IEEE Xplore. Restrictions apply.
0018-9456 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIM.2021.3115203, IEEE
Transactions on Instrumentation and Measurement
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2021 5
0 0.2 0.4 0.6 0.8 1
x (input)
0
0.2
0.4
0.6
0.8
1
y (output)
(a) Transfer
characteristic
0.1
0.297
0.742
(c) Output
0.2
0.365
0.088
(b) Input
Time
(t)
Time (t)
Fig. 7. Enhancement of pulsatile structure using the proposed transfer profile.
(a) Desired transfer profile, (b) input pulse waveform, and (c) output pulse
waveform.
envelope of the signal and further computing its instantaneous
energy. Finally, a transfer profile is proposed to emphasize the
pulsatile region while other regions are being suppressed. It
is illustrated by an example in Fig. 7. The desired profile can
be mathematically modeled using the following exponential
function, whose idea is inspired from the sigmoid logistic
function:
y=1ebx
1 + aebx ,x[0,1], a, b N(4)
where, aand bare the modelling parameters that are estimated
in a least square sense. The estimated parameters are indicated
by black dotted circle along with 95 % confidence interval
represented by black rectangular window in Fig. 8. Besides
modeled parameters, it illustrates a color-grid architecture
with the distributions of a and b parameters for representing
achieved correlations when the estimated transfer profile is
compared with the desired one. The ViCG-peaks are localized
in the form of sharp impulses and they are easily identified
using an amplitude-temporal-based thresholding scheme. Sub-
sequently, the cardiac intervals are measured from the detected
peaks using Eq. (2). The cardiac intervals measured from the
consecutive VP instants are statistically compared with the
cardiac cycles obtained from the reference ECG and PPG
signals. Despite having different signal acquisition mecha-
nisms, both ViCG and PPG implicitly measure arterial blood
flow at central and peripheral sites, respectively. Hence, the
PPG signals are also considered along with the ECG signals
for a comparative study. Cardiac intervals of the reference
ECG signal are calculated from the successive R-peaks, that
are detected using the well-known Pan-Tomkins algorithm.
Whereas, the PPG peaks, i.e., systolic peaks (SP) are used
to characterize the cardiac intervals in the PPG signals, and
they are detected using the algorithm presented in [29].
IV. RES ULTS AND DISCUSSION
To study the frequency characteristic for different patterns
of ViCG signals, spectrum analysis of the signal is carried out.
Fig. 9(a) shows various patterns of ViCG cycles collected from
0 10 20 30 40 50 60 70
"a" value
0
10
20
30
40
50
60
70
"b" value
0.7
0.8
0.9
b1a2
a1
a1: 36.12
a2: 41.73
b1: 15.7
b2: 16.26
b2
a: 38.92
b: 15.98
Maximum
correlation
point:
95 %
confidence
interval
Correlation
coefficient: 0.9999
Fig. 8. Estimation and validation of modeling parameters. Mesh plot
represents variations of correlation for achieved profiles corresponding to
various combinations of "a" and "b" parameters. Note that among all other
combinations, the estimated parameters indicated by black dotted circle yield
the maximum correlation.
(A) Different patterns of ViCG cycles
0 5 10 15 20
0
0.02
0.04
0.06
0.08
Power spectral density (nu2/Hz)
(B) Spectrum analysis of ViCG cycles
0 5 10 15 20
Frequency (Hz)
0
0.01
0.02
0.03
0.04
(b)
(c) (d)
(e) (f)
(a) Subject #1 Subject #2
Subject #3 Subject #4
Subject #5 Subject #6
(h)
(g) Power spectrums of
different ViCG cycles
Ensemble averaged
power spectrum
Fig. 9. Frequency domain analysis of ViCG signals. (a) ViCG cycles of six
individual subjects, (b) Power spectrum of corresponding ViCG cycles, and
(c) Ensembled power spectrum.
different individuals. Power spectrum is estimated for each of
the ViCG cycles as shown in Fig. 9(b), whereas (c) represents
the ensemble averaging of all the power spectrums. It is
observed that most of the signal power is distributed among the
frequencies that lies under 10 Hz. Thus, the spectrum analysis
confirms the frequency range of the ViCG signal.
To assess the cardiac cycle extraction, a beat-by-beat com-
parisons of our ViCG signals with reference ECG and PPG
signals were carried out by comparing the respective estimated
cardiac intervals. To detect the VP points, initially the zero
mean and amplitude normalized ViCG signal is pre-processed
using a Butterworth HPF of sixth order. Subsequently, pulsatile
profile of the ViCG is emphasized using the proposed envelope
construction scheme. Finally, VP locations are determined by
employing amplitude and temporal based thresholding scheme.
The peak detection results of all three signals and the com-
parisons of their derived cardiac intervals are illustrated by an
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY GUWAHATI. Downloaded on September 24,2021 at 07:09:25 UTC from IEEE Xplore. Restrictions apply.
0018-9456 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIM.2021.3115203, IEEE
Transactions on Instrumentation and Measurement
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2021 6
-0.2
0
0.2
0.4
PPG (n.u.)
-0.5
0
0.5
1
ViCG (n.u.)
600
800
1000
1200
RRint, VVint
RR intervals VV intervals
0 20 40 60 80 100 120
Time (s)
600
800
1000
1200
PPint, VV int
PP intervals VV intervals
-0.5
0
0.5
1
ECG (n.u.)
Detected peaks
(a)
(b)
(c)
(e)
(d)
Fig. 10. Experimental results for peak detection of cardiac signals and
comparison of their derived heart cycles. (a) Detected R peaks in ECG, (b)
detected SP peaks in PPG, (c) detected VP peaks in ViCG, (d) extracted
cardiac intervals from RR and VV (i.e. VP-VP) intervals, and (e) extracted
cardiac intervals from PP (i.e. SP-SP) and VV intervals. (Note that the
amplitude is represented in normalized unit (n.u.) for all the signals and
cardiac interval in ms.)
example in Fig. 10. The detected peaks, including R, SP, and
VP, are highlighted in Fig. 10(a)-(c) on their respective signals.
The detected VPs provide a precise idea about maximum blood
turbulence at the carotid. This fiducial information could be
utilized as a surrogate to assess the central blood flow at the
aorta. It can also help in rendering models for the assessment
of heart-power. The trace of cardiac intervals obtained from
the detected VP peaks is compared with the R-R and P-P
(PPG’s SP-to-SP interval) intervals as shown in Fig. 10(d) and
(e), respectively. It is qualitatively observed that the cardiac
intervals of both the signal pairs show similar traces. Thus,
the VP time-instants can be used to extract the physiological
parameters for cardiac health assessment. The comparison is
also evaluated for both the reference signals using the Bland-
Altman analysis. Fig. 11 shows the regression and Bland-
Altman plots for cardaic intervals of all the subjects. All the
measured values obtained from the plots are listed in Table II.
It is observed that the Pearson’s R-squared correlation coef-
ficient between the RR and VV (ViCG’s VP-to-VP interval)
intervals is 0.93, whereas it is found to be 0.92 between the
PP (PPG’s SP-to-SP interval) and VV intervals. Both exhibit
indeed strong correlations. Whereas their root mean square
errors (RMSE) are found very small, i.e., 28 ms each. Also,
most of the data lies inside the limits of agreement (LOA) as
shown by the dashed line in Bland-Altman plots. These results
show a good correlation and agreement of VV intervals with
both RR and PP intervals. This experiment also validates the
proposed signal for heart rate variability (HRV) analysis. Thus,
Fig. 11. Comparison of cardiac intervals in [1st row] ViCG-ECG and [2nd
row] ViCG-PPG pairs. Here, (a) and (c) represent correlation plots, while
(b) and (d) show Bland-Altman plots. Abbreviations used— VV: VP to VP
interval, RR: R to R interval, PP: SP to SP interval, MD: Mean difference,
and SD: Standard deviation.
TABLE II
PERFOR MANCE R ESU LTS FO R STATISTI CAL COM PARI SON OF VICG
DERIV ED CARD IAC I NTE RVALS
Pairs Regression Plot Bland-Altman Plot
RMSE (ms) r2y=ax+b MD LOA [L, U](ms)
(RR, VV) 28 0.93 1.00x+1.77 -0.01 -54, 54
(PP, VV) 28 0.92 0.99x+5.31 0.03 -55, 55
RMSE: Root mean square error, LOA: Limits of agreement, L and U represent lower
and upper LOA, respectively.
the information content of ViCG signal can be used along with
other cardiac signals for healthcare applications. As a wireless
stretchable skin sensor for the measurement of multi-cardiac
signals introduced in [30] for infants, the future directions of
our study could also be shaped with the investigation and
development of sensor-assembly for more compactness and
comfort.
V. CONCLUSION
In this study, a novel noninvasive approach, vibrocaroti-
dography, is proposed for cardiac health monitoring by mea-
suring and accessing central blood-flow information. This
paper mainly presents a pilot study on ViCG, its physiologi-
cal mapping and generation, spectro-temporal characteristics,
and possible clinical applications. The signal is acquired by
placing a miniaturized MEMS-based accelerometric sensor at
the common carotid on the neck surface. Rhythmic carotid
pulsations caused due to varying blood turbulence, originated
from aorta, are responsible for the ViCG signal generation.
The lower infrasonic ViCG signal is characterized by VP
peaks that reflect the information regarding maximum blood
turbulence instances at the carotid. The VP peaks are precisely
detected by employing the proposed envelope construction
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY GUWAHATI. Downloaded on September 24,2021 at 07:09:25 UTC from IEEE Xplore. Restrictions apply.
0018-9456 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIM.2021.3115203, IEEE
Transactions on Instrumentation and Measurement
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2021 7
technique, and subsequently, all the cardiac intervals are mea-
sured from the detected peaks. The performance is evaluated
by comparing the experimental results obtained from the
proposed signal with the reference ECG and PPG signal
measurements. A high similarity is qualitatively observed in
the cardiac interval pair traces, (i) R-R and V-V intervals, and
(ii) P-P and V-V intervals. The performance is validated by
performing the Bland-Altman analysis on them. The Pearson’s
R-squared correlation of 0.93 and 0.92 are obtained for the
cardiac intervals measured from both the ECG-ViCG and PPG-
ViCG signal pairs, respectively. Moreover, the analysis of
both the pairs reveals an approximately null bias with limits
of agreement upto 54 ms and 55 ms, respectively. It shows
quite a good agreement for both the cardiac signal pairs. The
performance results show the potential of the ViCG for heart
cycle extraction and HRV analysis. As a vital diagnostic tool,
the ViCG may be utilized in regular health-monitoring applica-
tions, such as heart rate estimation, assessment of central blood
flow variation, aortic pressure and turbulence, and detection
of pulse arrival time. It may also contribute in diagnosing
various cardiac pathological cases, such as higher or reduced
stroke volume, valvular defects like mitral deficiency, aortic
stenosis and regurgitation, arteriovenous fistula, and so on.
Thus, the ViCG signal may be employed for continuous
health monitoring in home, ambulatory, and hospital healthcare
scenarios. However, this article does not cover any extensive
and comprehensive experimentations to validate its utility
for pathological applications. We expect more explorations,
investigations, and validations from cardiac researchers to
establish the multimodal diagnosis in this domain.
ACKNOW LE DG ME NT S
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
The authors would like to thank the anonymous reviewers for
their useful suggestions and comments.
REFERENCES
[1] E. A. Rosei, G. Mancia, M. F. O’Rourke, M. J. Roman, M. E. Safar, H.
Smulyan, J. G. Wang, I. B. Wilkinson, B. Williams, and C. Vlachopou-
los, “Central blood pressure measurements and antihypertensive therapy:
a consensus document," Hypertension, vol. 50, no. 1, pp. 154–160, 2007.
[2] J. A. Chirinos, J. P. Zambrano, S. Chakko, A. Veerani, A. Schob, H. J.
Willens, G. Perez, and A. J. Mendez, “Aortic pressure augmentation pre-
dicts adverse cardiovascular events in patients with established coronary
artery disease," Hypertension, vol. 45, no. 5, pp. 980–985, 2005.
[3] C. M. McEniery, N. Yasmin, B. McDonnell, M. Munnery, S. M. Wallace,
C. V. Rowe, J. R. Cockcroft, and I. B. Wilkinson, “Central pressure:
variability and impact of cardiovascular risk factors: the Anglo-Cardiff
Collaborative Trial II," Hypertension, vol. 51, no. 6, pp. 1476–1482,
2008.
[4] M. J. Roman, R. B. Devereux, J. R. Kizer, E. T. Lee, J. M. Galloway,
T. Ali, J. G. Umans, and B. V. Howard, “Central pressure more strongly
relates to vascular disease and outcome than does brachial pressure: the
Strong Heart Study," Hypertension, vol. 50, no. 1, pp. 197–203, 2007.
[5] M. E. Safar, J. Blacher, B. Pannier, A. P. Guerin, S. J. Marchais, P. M.
Guyonvarc’h, and G. M. London, “Central pulse pressure and mortality
in end-stage renal disease," Hypertension, vol. 39, no. 3, pp. 735–738,
2002.
[6] B. Williams, P. S. Lacy, S. M. Thom, K. Cruickshank, A. Stanton, D.
Collier, A. D. Hughes, H. Thurston, M. O’Rourke, CAFÉ investigators,
and Anglo-Scandinavian Cardiac Outcomes Trial Investigators, “Differ-
ential impact of blood pressure-lowering drugs on central aortic pressure
and clinical outcomes: principal results of the Conduit Artery Function
Evaluation (CAFE) study," Circulation, vol. 113, no. 9,pp. 1213–1225,
2006.
[7] S. Casaccia, E. J. Sirevaag, E. J. Richter, J. A. O’Sullivan, L. Scalise,
and J. W. Rohrbaugh, “Features of the non-contact carotid pressure
waveform: Cardiac and vascular dynamics during rebreathing," Review
of Scientific Instruments, vol. 87, no. 10, pp. 102501, 2016.
[8] J. J. VITEK, “Femoro-cerebral angiography: analysis of 2,000 consec-
utive examinations, special emphasis on carotid arteries catheterization
in older patients," American Journal of Roentgenology, vol. 118, no. 3,
pp. 633–647, 1973.
[9] D. Z. Zambrana, J.M. Vicente-Samper, C. G. Juan, V. E. Sala, and J.
M. Sabater-Navarro, “Non-Invasive Device for Blood Pressure Wave
Acquisition by Means of Mechanical Transducer," Sensors, vol. 19, no.
19, pp. 4311, 2019.
[10] M. F. O’Rourke, “Carotid artery tonometry: pros and cons," American
Journal of Hypertension, vol. 29, no. 3, pp. 296–298, 2016.
[11] B. Spronck, T. Delhaas, J. O. Roodt, and K. D. Reesink, “Carotid
artery applanation tonometry does not cause significant baroreceptor
activation," American Journal of Hypertension, vol. 29, no. 3, pp. 299–
302, 2016.
[12] M. Takino, Y. Takino, and K. Sugahara, “Apparatus and method for
measurement of digital pressure applied to carotid sinus for causing
carotid sinus syndrome," Acta neurovegetativa, vol. 26, no. 1, pp. 93–
103, 1964.
[13] C. C. Chang, P. H. Tsui, C. C. Chang, K. J. Chang, C. L. Wang, J.
J. Hwang, C. C. Chu, C. N. Chen, L. Y. Lin, and inventors; National
Taiwan University, assignee, “Carotid pulse measurement device," U.S.
patent 11/783 447, Jun. 26, 2008.
[14] A. C. Rossi, P. J. Brands, and A. P. Hoeks, “Automatic recognition of
the common carotid artery in longitudinal ultrasound B-mode scans,"
Medical image analysis, vol. 12, no. 6, pp. 653–665, 2008.
[15] T. Idzenga, S. Holewijn, H. H. Hansen, and C. L. Korte, “Estimating
cyclic shear strain in the common carotid artery using radiofrequency
ultrasound," Ultrasound in medicine and biology, vol. 38, no. 12, pp.
2229–2237, 2012.
[16] S. J. Vermeersch, E. R. Rietzschel, M. L. Buyzere, D. D. Bacquer, G. D.
Backer, L. M. Van Bortel, T. C. Gillebert, P. R. Verdonck, and P. Segers,
“Determining carotid artery pressure from scaled diameter waveforms:
comparison and validation of calibration techniques in 2026 subjects,"
Physiological measurement, vol. 29, no. 11, pp. 1267, 2008.
[17] J. Kips, F. Vanmolkot, D. Mahieu, S. Vermeersch, I. Fabry, J. D Hoon, L.
V. Bortel, and P. Segers, “The use of diameter distension waveforms as
an alternative for tonometric pressure to assess carotid blood pressure,"
Physiological measurement, vol. 31, no. 4, pp. 543, 2010.
[18] J. Luo, R. X. Li, and E. E. Konofagou, “Pulse wave imaging of the
human carotid artery: an in vivo feasibility study," IEEE transactions
on ultrasonics, ferroelectrics, and frequency control, vol. 59, no. 1, pp.
174–181, 2012.
[19] C. Wang, X. Li, H. Hu, L. Zhang, Z. Huang, M. Lin, Z. Zhang, Z.
Yin, B. Huang, H. Gong, and S. Bhaskaran, “Monitoring of the central
blood pressure waveform via a conformal ultrasonic device," Nature
biomedical engineering, vol. 2, no. 9, pp. 687–695, 2018.
[20] D. Buxi, J. M. Redouté, and M. R. Yuce, “Cuffless blood pressure
estimation from the carotid pulse arrival time using continuous wave
radar," in Proc. IEEE Engineering in Medicine and Biology Society,
2015, pp. 5704–5707.
[21] L. Casacanditella, et al., “Indirect measurement of the carotid arterial
pressure from vibrocardiographic signal: Calibration of the waveform
and comparison with photoplethysmographic signal," in Proc. IEEE
Engineering in Medicine and Biology Society, 2016, pp. 3568–3571.
[22] V. Mancini, D. Tommasin, Y. Li, J. Reeves, R. Baets, S. Greenwald,
P. Segers, and CARDIS consortium, “Detecting carotid stenosis from
skin vibrations using Laser Doppler Vibrometry–An in vitro proof-of-
concept," PloS one, vol. 14, no. 6, pp. e0218317, 2019.
[23] S. Casaccia, E. J. Sirevaag, E. J. Richter, J. A. O’Sullivan, L. Scalise,
and J. W. Rohrbaugh, “Features of the non-contact carotid pressure
waveform: Cardiac and vascular dynamics during rebreathing," Review
of Scientific Instruments, vol. 87, no. 10, pp. 102501, 2016.
[24] L. Antognoli, S. Moccia, L. Migliorelli, S. Casaccia, L. Scalise, and
E. Frontoni, “Heartbeat Detection by Laser Doppler Vibrometry and
Machine Learning," Sensors, vol. 20, no. 18, pp. 5362, 2020.
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY GUWAHATI. Downloaded on September 24,2021 at 07:09:25 UTC from IEEE Xplore. Restrictions apply.
0018-9456 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIM.2021.3115203, IEEE
Transactions on Instrumentation and Measurement
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2021 8
[25] A. D. Kaplan, J. A. O’Sullivan, E. J. Sirevaag, S. D. Kristjansson,
P. H. Lai, and J. W. Rohrbaugh, “Hidden state dynamics in laser
Doppler vibrometery measurements of the carotid pulse under resting
conditions," in Proc. IEEE Engineering in Medicine and Biology,, 2010,
pp. 5273–5276.
[26] T. Choudhary, L.N. Sharma, and M.K. Bhuyan, “Method and Tech-
nology for Accelerometric Signal Recording of a Novel Vibrocaroti-
dogram (ViCG) with Seismocardiogram (SCG)," Indian Patent Request
202031026802, 2020.
[27] T. Choudhary, M.K. Bhuyan, and L.N. Sharma, “Delineation and
Analysis of Seismocardiographic Systole and Diastole Profiles," IEEE
Transactions on Instrumentation and Measurement, vol. 70, Art no.
4000108, pp. 1–8, 2021.
[28] Accessed on May. 2021 [online] Available:
https://www.entclinic.com.au/vascular-anatomy-of-the-neck/
[29] T. Choudhary and M. S. Manikandan, “Robust photoplethysmographic
(PPG) based biometric authentication for wireless body area networks
and m-health applications," in Proc. IEEE National Conference on
Communication (NCC), 2016, pp. 1–6.
[30] S. Xu et al., “Wireless skin sensors for physiological monitoring of
infants in low-income and middle-income countries," The Lancet Digital
Health, vol. 3, no. 4, pp. e266–e273, 2021.
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY GUWAHATI. Downloaded on September 24,2021 at 07:09:25 UTC from IEEE Xplore. Restrictions apply.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Globally, neonatal mortality remains unacceptability high. Physiological monitoring is foundational to the care of these vulnerable patients to assess neonatal cardiopulmonary status, guide medical intervention, and determine readiness for safe discharge. However, most existing physiological monitoring systems require multiple electrodes and sensors, which are linked to wires tethered to wall-mounted display units, to adhere to the skin. For neonates, these systems can cause skin injury, prevent kangaroo mother care, and complicate basic clinical care. Novel, wireless, and biointegrated sensors provide opportunities to enhance monitoring capabilities, reduce iatrogenic injuries, and promote family-centric care. Early validation data have shown performance equivalent to (and sometimes exceeding) standard-of-care monitoring systems in premature neonates cared for in high-income countries. The reusable nature of these sensors and compatibility with low-cost mobile phones have the future potential to enable substantially lower monitoring costs compared with existing systems. Deployment at scale, in low-income countries, holds the promise of substantial improvements in neonatal outcomes.
Article
Full-text available
Background: Heartbeat detection is a crucial step in several clinical fields. Laser Doppler Vibrometer (LDV) is a promising non-contact measurement for heartbeat detection. The aim of this work is to assess whether machine learning can be used for detecting heartbeat from the carotid LDV signal. Methods: The performances of Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF) and K-Nearest Neighbor (KNN) were compared using the leave-one-subject-out cross-validation as the testing protocol in an LDV dataset collected from 28 subjects. The classification was conducted on LDV signal windows, which were labeled as beat, if containing a beat, or no-beat, otherwise. The labeling procedure was performed using electrocardiography as the gold standard. Results: For the beat class, the f1-score (f1) values were 0.93, 0.93, 0.95, 0.96 for RF, DT, KNN and SVM, respectively. No statistical differences were found between the classifiers. When testing the SVM on the full-length (10 min long) LDV signals, to simulate a real-world application, we achieved a median macro-f1 of 0.76. Conclusions: Using machine learning for heartbeat detection from carotid LDV signals showed encouraging results, representing a promising step in the field of contactless cardiovascular signal analysis.
Article
Full-text available
Blood pressure wave monitoring provides interesting information about the patient's cardiovascular function. For this reason, this article proposes a non-invasive device capable of capturing the vibrations (pressure waves) produced by the carotid artery by means of a pressure sensor encapsulated in a closed dome filled with air. When the device is placed onto the outer skin of the carotid area, the vibrations of the artery will exert a deformation in the dome, which, in turn, will lead to a pressure increase in its inner air. Then, the sensor inside the dome captures this pressure increase. By combining the blood pressure wave obtained with this device together with the ECG signal, it is possible to help the screening of the cardiovascular system, obtaining parameters such as heart rate variability (HRV) and pulse transit time (PTT). The results show how the pressure wave has been successfully obtained in the carotid artery area, discerning the characteristic points of this signal. The features of this device compare well with previous works by other authors. The main advantages of the proposed device are the reduced size, the cuffless condition, and the potential to be a continuous ambulatory device. These features could be exploited in ambulatory tests.
Article
Full-text available
Early detection of asymptomatic carotid stenosis may help identifying individuals at risk of stroke. We explore a new method based on laser Doppler vibrometry (LDV) which could allow the non-contact detection of stenosis from neck skin vibrations due to stenosis-induced flow disturbances. Experimental fluid dynamical tests were performed with water on a severely stenosed patient-specific carotid bifurcation model. Measurements were taken under various physiological flow regimes both in a compliant and stiff-walled version of the model, at 1 to 4 diameters downstream from the stenosis. An inter-arterial pressure catheter was positioned as reference. Increasing flow led to corresponding increase in power spectral density (PSD) of pressure and LDV recordings in the 0–500 Hz range. The stiff model lead to higher PSD. PSD of the LDV signal was less dependent on the downstream measurement location than pressure. The strength of the association between PSD and flow level, model material and measuring location was highest in the 0–50 Hz range, however useful information was found up to 200 Hz. This proof-of-concept suggests that LDV has the potential to detect stenosis-induced disturbed flow. Further computational and clinical validation studies are ongoing to assess the sensitivity and specificity of the technique for clinical screening.
Article
Full-text available
Continuous monitoring of the central blood pressure waveform from deeply embedded vessels such as the carotid artery and jugular vein has clinical value for the prediction of all-cause cardiovascular mortality. However, existing non-invasive approaches, including photoplethysmography and tonometry, only enable access to the superficial peripheral vasculature. Although current ultrasonic technologies allow non-invasive deep tissue observation, unstable coupling with the tissue surface resulting from the bulkiness and rigidity of conventional ultrasound probes introduces usability constraints. Here, we describe the design and operation of an ultrasonic device that is conformal to the skin and capable of capturing blood pressure waveforms at deeply embedded arterial and venous sites. The wearable device is ultrathin (240 μm) and stretchable (with strains up to 60%), and enables the non-invasive, continuous and accurate monitoring of cardiovascular events from multiple body locations, which should facilitate its use in a variety of clinical environments. © 2018, The Author(s), under exclusive licence to Springer Nature Limited.
Conference Paper
Full-text available
Ambulatory blood pressure monitors based on pulse transit time are limited by the challenge of changing vascular tone. This study focuses on the use of the carotid artery as an alternative location for arterial pulse acquisition. We use continuous wave radio frequency (RF) radar coupled directly to the body to detect the pulse wave signal. We have shown that the blood pressure-pulse transit time calibration using the carotid pulse is as accurate as that of the radial arterial pulse. The results of this investigation may be useful in developing wearable sensors for long-term monitoring of the pulse wave signal at the carotid artery.
Conference Paper
Full-text available
In this paper, we present noise-robust photoplethys-mographic (PPG) based biometric authentication method for wireless body area networks and m-health applications. The method consists of four steps: (i) preprocessing of PPG signals, (ii) systolic peak detection, (iii) ensemble averaged pulsatile waveform extraction and (iv) pulsatile waveform similarity matching using a normalized cross correlation (NCC) measure. The performance of the proposed method is tested and validated using different types of PPG signals taken from the standard PPG databases. For predefined threshold of 0.997, the NCC-based PPG biometric method achieves an average false rejection rate (FRR) of 0.32 and false acceptance rate (FAR) of 0.32. Performance evaluation results show that the proposed method achieves consistent authentication results as compared to the other methods under different kinds of artifacts and noise.
Conference Paper
The detection of arterial Blood Pressure waveform provides important information about the subject health status. Laser Doppler Vibrometry (LDV) is a non-contact technique with high sensitivity able to detect mechanical movements of the arterial wall; several previous studies have shown that LDV is able to characterize cardiac activity. Photoplethysmogram (PPG) quantifies the digital volume artery pulse, which has been demonstrated to be closely related to the pressure signal measured by an arterial tonometer. In this paper, an indirect measurement of carotid arterial pressure by means of LDV is presented. Moreover, a comparison between LDV and PPG is conducted in order to estimate the time interval between opening and closing of the aortic valve, that is the Left Ventricular Ejection Time (LVET). Results show an average reduction of around 20% of the systolic pressure derived from LDV signal measured over the carotid artery with respect to the systolic pressure measured at brachial level (i.e. peripheral pressure value). Finally, the comparison between LDV and PPG in the estimation of LVET shows a mean percentage deviation <10%. So, in conclusion, it can be stated that LDV technique has the potential of providing a displacement waveform that, adequately calibrated, can furnish significant information about pressure waveform.