AUTOMATIC EVALUATION OF THE
QUANTITATIVE SEISMOCARDIOGRAM
Z. Trefny, J. Svacinka, S. Trojan, J. Slavicek, P. Smrcka and M. Trefny
Cardiological Laboratory, U Průhonu 52, Prague 7, Czech Republic
Institute of Physilogy, 1
st
Medical Faculty, Charles University in Prague, Albertov 5, Prague 2, Czech Republic
Faculty of Biomedical Engineering, Czech Technical University in Prague, Studnickova 7, Prague 2, Czech Republic
Keywords: Time-domain segmentation of the seismocardiogram, J-wave recogniton.
Abstract: The device for quantitative seismocardiography (Q-SCG) detects cardiac vibrations caused by the heart
activity, the measuring sensor is usually placed in the plate of the chair – additional instruments applied on
the proband’s body are not required. The results of the Q-SCG analysis are usable in various clinical fields.
The first and most important step in the process of detection of significant characteristics of measured Q-
SCG curves is to detect pseudo-periods in the signal regardless of the initial pseudo-period position. Other
characteristics can be acquired by a relatively simple process over the appointed pseudo-period. The
experimental equipment for the Q-SCG measuring and analysis was developed and also special algorithms
for preprocessing, segmentation and interactive analysis of the Q-SCG signal were developed. In this
contribution technical principles of the quantitative seismocardiography are introduced; the method is easy,
robust and is appropriate for real-time Q-SCG processing.
1 INTRODUCTION
Ballistocardiography (BCG): In 1936, Starr began
the era of high-frequency ballistocardiography,
which lasted approximately 15 years. Different types
of instruments were developed, on which the
displacement, velocity or acceleration of the body
lying on a table was measured. Later studies showed
that there are difficulties when comparing records
registered on different apparatuses. This was mainly
caused by two factors: (a) the instrument’s natural
frequency, (b) the instrument’s damping.
Figure: 1: Records registered using the old BCG
instrument with a frequency of 2Hz and critical damping.
The lower curve depicts the effect of force applied, which
is of the same intensity but differs in the duration. The
upper curve is a record, from which one cannot determine
either size or duration of the acting force.
Quantitative ballistocardiography (Q-BCG): Fo-
llowing the critical evaluation of all these facts, we
began in 1952 our own experiments related to the
construction of an apparatus which would lack the
aforementioned shortcomings. We constructed an
apparatus whose advantages lie not only in the
simplicity of its design, but also in its important
functional qualities. To achieve a minimal distortion
caused by the transmission from the origin of the
force to the recorder it is necessary that the natural
frequencies of the transmission systems are as far as
possible from the mentioned frequency range.
The cardiovascular activity is manifested by a
force acting on the human body which represents a
mechanical vibratory system transmitting the force
to the balistocardiographic apparatus.
The basic part of our portable quantitative
balistocardiograph is a very rigid piezoelectric force
transducer resting on a rigid chair. The examined
person sits (Figure 2) on the light seat placed on the
transducer and the force caused by the
cardiovascular activity is measured in this way. The
output of the piezoelectric pick-up is fed into an
operational amplifier.
463
Trefny Z., Svacinka J., Trojan S., Slavicek J., Smrcka P. and Trefny M. (2007).
AUTOMATIC EVALUATION OF THE QUANTITATIVE SEISMOCARDIOGRAM.
In Proceedings of the Ninth International Conference on Enterprise Information Systems - AIDSS, pages 463-467
DOI: 10.5220/0002412604630467
Copyright
c
SciTePress
Figure 2: Position of the examined person during the
QBCG session.
The advantage of the piezoelectric transducer is
in very low compliance together with a very high
natural frequency. Another advantage of the rigid
pick-up is the fact that it can be preloaded with a
substantial static force – the weight of the examined
person, and it is still possible to measure the
alternating forces of the magnitude of g+ (gram as
weight). The measured force is registered (REG)..
Figure 3: Records registered using the QBCG instrument.
The lower curve depicts the effect of force applied. The
upper curve is a record, from which we can determine size
and duration of the acting force. Compare with BCG
record on Figure 1.
Quantitative seismocardiography: (Q-SCG: This
method enables the recording of force applied
without phase or time deformation. Thus, heart rate
may be monitored and analysed using the method of
heart rate variability: statistical and autocorrelation
analysis, spectral analysis, total effect of regulation,
vegetative homeostasis, activity of subcortical
centre, activity of the vasomotor centre and stress
index (SI). For example the SI we can calculate
simply as
SI = AMo / (2 . Mo . MxDMn)
where Mo is modus of the RR interval series, AMo
is the amplitude of the modus and MxDMn is a
difference between maximal and minimal RR
interval . The SI describes the tension of regulative
systems, represents the degree of stress of regulatory
systems (the degree of the predominance of the
activity of the central mechanism of regulation over
the autonomic ones).
The method of Q-SCG was designated by the
laboratory employees as an absolutely non-invasive,
and the persons examined did not have any
electrodes attached to the body surface and were not
connected by cables to the registering instrument.
This new field of monitoring heart activity, whereby
we determine both amplitude-force and time-
frequency relationships, is termed quantitative
seismocardiography. Thus, one may determine the
force-response of the cardiovascular system to
changes in external stimuli, as well as the
autonomous nervous system regulation of the
circulation and the activity of the sympathetic and
parasympathetic systems.
2 METHODS OF QSCG
MEASUREMENT AND
ANALYSIS
2.1 Experimental Equipment
In terms of practical use, a portable telemetric
system for the Q-SCG measuring has been
developed. This system allows data to be acquired
and assessed using quantitative seismocardiography
(Q-SCG). It is composed of the three HW modules
that are telemetrically interconnected with the option
of interconnecting through a metallic line. These are
the seismocardiographic, the accelerometric modules
and a module for the data transfer interconnected
with the PC through the USB interface. Block
scheme of the whole system is on figure 6.
Figure 4: Main sensor of she Q-SCG measuring equipment
- detail of the solid-state accelerometer between measuring
metal plates.
ICEIS 2007 - International Conference on Enterprise Information Systems
464
Sensing mechanical body reactions, which are
induced in response to the cardiovascular dynamics,
is provided by the seismocardiographic module,
which reads the strain coming from the mechanical
deformation of the piezo-electric plate. This sensing
module (figures 4 and 5) is mounted on a special
device, which works by transmitting the mechanical
body reactions onto the piezo-electric element. The
data transfer module is designed to transmit the data
from the radio-module into the PC through the USB
interface.
Figure 5: Measuring plates of the proposed Q-SCG device.
Figure 6: Block scheme of the experimental Q-SCG
device.
2.2 Algorithm for the Time-Domain
Segmentation of the Q-SCG
Algorithms for preprocessing, segmentation and
interactive analysis of the Q-SCG signal were
developed. In this contribution we will focuse on the
original method for basic segmentation of the Q-
SCG signal in time-domain; this first step is crucial
for the successfulness of the whole diagnostic
process. Our method is relatively simple and was
developed for the detection of Q-SCG pseudo-
periods in real time. The method is derived from a
well-known and robust algorithm for QRS complex
detection in traditional electrocardiograms (ECG),
originally developed by Hamilton et al. The
algorithm was based on the first derivative of the
input signal and many thresholds and parameters are
automatically adapted to individual changes in the
input signal using sophisticated empirical rules. The
results (position of the dominant – so-called R -
wave) are obtained with some detection delay
(above 200 ms). For details on the algorithm, see
(Hamilton).
For our purposes it is important that the initial
values of many parameters are adjustable and by
modification of these values the original method was
slightly adapted to Q-SCG’s different curve
morphology. Namely the following parameters were
changed: (1) length of the first derivative from the
original 10 ms to 80 ms, (2) length of the high-pass
pre-filter from 125 ms to 350 ms, (3) length of
moving window integration from 80 ms to 200 ms.
Optimal values were selected experimentally in
order to achieve the best detection results.
Additionally, we developed a special backward
searching process for the precise detection of the
position of the I-wave and J-wave in each Q-SCG
pseudo-period.
The function of the whole algorithm is as follows:
output of the traditional ECG QRS detector gives the
rough position of the systolic complex inside the Q-
SCG - candidate X. Then the specific morphology of
the Q-SCG curve is utilized to backward search the
position of the J-wave – we expect the first big
negative peak in MTI samples (about 100 ms). If the
detection is successful, we assign the position of the
peak as the I-wave; see Figure 7.
Finally we search forward for the position of the
J-wave, which we expect to be the first big positive
peak in maximally MTJ samples (about 160 ms), see
Figure 8.
Control
unit,
A/D
Converter
Commu-
nication
interface
WIFI
Bluetooth
Q
SCG unit
ECG measuring
unit
Auxiliary
measuring units
PATIENT‘S MEASURING DEVICE
PATIENT
PC
WORKSTATION
FOR
SUPERVISON,
DATA
PROCESSING
WIRELESS CONNECTION
AUTOMATIC EVALUATION OF THE QUANTITATIVE SEISMOCARDIOGRAM
465
For the peak-detection we used a very simple
method based on the first difference (length 15 ms):
when the transition from negative to positive value
of the difference occurs, then the sequence is marked
as a negative peak; the transition from a positive to
negative difference means a positive peak. If
searching for the J-wave or the I-wave fails,
candidate “X” is rejected and the algorithm
continues without detection of the Q-SCG pseudo-
period.
The rejection of “candidate X” is very important
step and it increases robustness of the whole
detection procedure against the artifacts – see
demonstration on the Figure 9.
The false detection of the dominant “candidate X”,
which is not a true Q-SCG cycle, was corrected by
the proposed simple backward searching algorithm,
because the morphology in the nearest neighborhood
of the point X3 does not match the detection rules –
backward searching for the I-wave in MTI samples
was not successful, the false positive detection of the
systolic complex was correctly rejected.
Our experimental software allows also automatic
extraction of classical Q-SCG hemodynamical
parameters, especially so called systolic force (F).
The current version of the system is designed for OS
Windows XP and has user-friendly interface. Block
scheme of the system is on figure 11.
3 CONCLUSION
For high-quality measurements we can obtain good-
looking signals for which both methods exhibit
excellent results. For disruptive and spurious signals
there is still a good chance of obtaining authentic
information because we first detect the impairments
and remove the particular interval of the signal.
For good-looking and typical signals, the
methods behave very well, achieving nearly
complete success (see Figure 12). The Q-SCG signal
offers specific information about functional
changes of the cardiovascular system regulation
which preceded the structural changes coming later.
The equipment is ready for use, algorithms for
automatic analysis of the Q-SCG signal are
prepared.
Quantitative seismocardiography probably offers
a more complex view of both inotropic and
chronotropic hearth function. It will be suitable for:
6000
7000
8000
9000
10000
11000
12000
13000
1 58 115 172 229 286 343 400 457
time [sam ples]
Force (quant. levels )
X
I
H
K
L
M
N
I
max
MTI
Figure 7: Backward local I-peak searching in the Q-
SCG cycle.
6000
7000
8000
9000
10000
11000
12000
13000
1 58 115 172 229 286 343 400 457
time [samples]
Force (quant. levels)
I J
H
K
L
M
N
J
max
MTJ
Figure 8: Forward local J-peak searching in the Q-SCG
cycle.
8800
10800
12800
1 11 21 31 41 51 61 71 81 91 101 111 121 131 141 151 161 171 181 191 201 211 221
time (samples)
Force (quant. levels)
X3
I
max
MTI
Figure 9: Rejection of the false beat detection. We searc
h
backward from “candidate X3” for the first big negative
peak. The I-wave must be recognized in MTI samples
(about 100 ms), so in this case the detection was not
successful.
5000
7000
9000
11000
13000
15000
1 73 145 217 289 361 433 505 577 649 721 793 865 937 1009 108 1 1153 1225
time (samples)
Force (quant. levels)
J
J
J
X1
X2 X3 X4
Figure 10: Result of the whole detection: false candidate
X3 was correctly rejected.
ICEIS 2007 - International Conference on Enterprise Information Systems
466
examining operators exposed to stress; for assessing
the effect of work, fatigue, mental stress; for
monitoring persons as part of disease prevention; for
determining a person’s ability to carry out their
duties both on the ground and in the air.
ACKNOWLEDGEMENTS
This work has been supported by the Ministry of
Education of the Czech Republic under project No.
MSM6840770012 and by the project EUREKA
E!3031.
REFERENCES
Jerosch-Herold, M. et al., 1999. The seismocardiogram as
magnetic-field-compatible alternative to the
electrocardiogram for cardiac stress monitoring, In
International Journal of Cardiac Imaging, 15(6), pp.
523-31
Trefny, Z. et al., 1996: Some physical aspects in
cardiovascular dynamics, In J. Cardiovasc. Diagnosis
and Procedures, 13(2), pp. 141 - 145
Hamilton, P. – Tompkins, W.J. (1987): Quantitative
investigation of QRS detection rules using the
MIT/BIH arrhythmia database, In IEEE Trans.
Biomed.Eng., 33, pp. 1158-65
Trefny Z. – David E. - Bayevsky R.M.: Achievements in
Space Medicine into Healt Care Practice and Industry,
Development of Space Cardiology metods in the
Earth's Health Service, Berlin 2001
Freisen G., Jannet T.: A comparison of the Noise
Sensitivity of Nine QRS Detection Algorithms’, IEEE
Trans. Biomed. Eng., 1990, 85(1)
Measuring
software
Visualization,
printing and
archivation
Setup and
calibration
Signal
acquisition
and data
Data processing unit
Unit for time-
domain
segmentation of the
Unit for extraction
of the
hemod
namical
Unit for extraction of
the HRV parameters
Figure 11: Block scheme of the software system. Presente
d
algorithm is in the box „Unit for time-domain segmentatio
n
of the Q-SCG curves“.
4000
6000
8000
10000
12000
14000
1 70 139 208 277 346 415 484 553 622 691 760 829 898 967 1036 1105 1174 1243 1312 1381 1450
Force (quant. levels)
time (samples)
J J
J
J J
I
I
I
I
I
Figure 12: Typical Q-SCG signal with correctly placed
reference points.
AUTOMATIC EVALUATION OF THE QUANTITATIVE SEISMOCARDIOGRAM
467