Abstract —A measurement system for heart sounds was
implemented by using ElectroMechanical Film (EMFi). Heart sounds are produced by the vibrations of the cardiac structure. An EMFi transducer attached to the skin of the chest wall converts the mechanical vibrations into an electrical signal. Furthermore, the signal is amplified and transmitted to the computer. The data is analyzed with Matlab ® software. The low-frequency components of the measured signal (respiration and pulsation of the heart) are filtered out as well as the 50 Hz noi. Also the power spectral density (PSD) plot is computed. In test measurements, the signal was measured with respiration and by holding breath. From the filtered signal, the first (S1) and the cond (S2) heart sound can be clearly en in both cas. In addition, from the raw data signals the respiration frequency and the heart rate can be determined. In future applications, with the EMFi material it is possible to implement a plaster-like transducer measuring vital signals.
I. I NTRODUCTION
HE auscultation of sounds produced by the human heart is a common procedure of medical practitioners. It is one of the fundamental tools in the diagnosis of heart dias [1] and it provides a lot of valuable information concerning the cardiovascular system [2].
Heart sounds are produced by the biological membranes of the heart when an event such as the opening or closing of a valve, vibration of the cardiac structure or acceleration or deceleration of blood occurs [3]. Besides the heart sounds, also murmurs may exist. Murmurs are vibrations due to the blood turbulence [4].
The vibrations caud by the heart sounds and murmurs travel to the chest wall where they can be heard as sounds [5]. The human ear is poorly suited for cardiac auscultation [1] due to the low amplitude and frequency content of the sounds, and thus a stethoscope is ud to transmit the sounds from the chest wall to the ear. The interpretation of the heart sounds is a special skill and it is mainly bad on the experti of the doctors [1], [6]. Hence it is more efficient to display the information provided by the heart sounds visually [3].
The phonocardiography (PCG) is an effective method for
Manuscript received April 2, 2007.
S. Kärki is with Measurement and Information Technology, Tampere University of Technology, P.O. Box 692, Tampere, FI-33101 Finland (corresponding author to provide phone: +358-3-31153573; fax: +358-3-31152171;e-mail:*****************).
M. Kääriäinen is with Department of Plastic Surgery, Tampere University Hospital, P.O. Box 2000, Tampere, FI-33521 Finland (e-mail: ************************).
J. Lekkala is with Measurement and Information Technology, Tampere University of Technology, P.O. Box 692, Tampere, FI-33101 Finland (e-mail:********************).
visual display of the heart sounds [3]. It is a uful diagnostic tool showing the timings and relative intensities of the heart sounds with graphic recordings and thus revealing information that the human ear cannot [7]. Phonocardiogram is obtained by positioning a microphone on the skin, and it requires a conversion of mechanical vibration into an electrical signal [8]. Since the heart sounds have low amplitudes, extraneous nois must be minimized in the vicinity of the patient [4].最好的作文
Various techniques of measuring the heart sounds have been reported. Most of the are bad on the phonocardiography. The PCG signal is measured, for example, by using a digital stethoscope, a microphone or a piezoelectric polymer film (polyvinylidene fluoride, PVDF) [3], [6], [9].
In this study, the heart sounds were measured with a transducer implemented by using ElectroMechanical Film (EMFi TM , EMFi is a registered trademark of Emfit Ltd). EMFi is a thin and elastic cellular polypropylene film consisting of three layers: smooth and homogenous surface layers
and a dominant, thicker midction full of flat voids parated by polypropylene layers [10]. The thickness of the material is only a few dozens of micrometers.
The EMFi material is nsitive to dynamic forces exerted normal to its surface. The nsor operation has a capacitive nature and it is bad on the thickness variations in the midction of the film caud by an external force [10]. Wide spectrum of applications and low ba material cost are the main advantages of EMFi. The material is uful especially in physiological measurements, where the material is in contact with skin or clothing. More information concerning the EMFi material can be found from references [10] and [11].
The structure of this paper is following. In Section 2, the physiological origin of the heart sounds is described. Section 3 introduces the measurement system. In Sections 4 and 5, the test measurements carried out and the results obtained are reported and discusd, respectively. Section 6 reprents the conclusions of this study.
II. H EART S OUNDS
The human heart is divided into the right and left halves, each consisting of an atrium and a ventricle. Between the atrium and ventricle in each half are the atrioventricular (AV) valves, called th
e tricuspid and mitral valves [12]. The opening and closing of the AV valves is a passive process resulting from pressure differences across the valves. The opening of the right ventricle into the pulmonary circulation
Measurement of heart sounds with EMFi transducer
Satu Kärki, Minna Kääriäinen and Jukka Lekkala
T
Proceedings of the 29th Annual International Conference of the IEEE EMBS Cité Internationale, Lyon, France August 23-26, 2007.
FrA06.4
and left ventricle into the aorta are also regulated by the valves, called the pulmonary and aortic valves [12].
The cardiac cycle consists of two phas: the period of ventricular contraction and blood ejection, systole, and ventricular relaxation, diastole, during which the ventricle fills with blood [12]. The sound
emitted by a human heart during a single cardiac cycle consists of two dominant events, known as the first heart sound S1 and the cond heart sound S2 [13]. The first heart sound depicts the ont of systole, and it is produced by the closure of mitral and tricuspid valves [14]. The cond heart sound is produced by the closure of the aortic and pulmonary valves, and it marks the end of systole and the beginning of diastole [14]. S1 and S2 are always audible in a normal patient [1].包粽子的过程
Additional heart sounds may occur in early diastole (the third heart sound, S3) and late diastole (the fourth heart sound, S4) [14]. S3 is sometimes audible in normal patients, but S4 is usually abnormal [14]. S3 and S4 usually cannot be heard with a stethoscope becau of their low frequency content [5].
In the auscultation of heart, different sound components may be audible. The primary components are the four heart sounds. The cond class is the murmurs [1] caud by the turbulence in blood flow [4]. Murmurs may occur in normal heart or they may be caud by structural abnormalities [1]. In addition, clicks and snaps may be heard during the cardiac cycle. The sounds indicate abnormality and they are associated with valve opening [1].
Sound consists of vibrations in different frequencies [5] and thus different heart disorders generate th
e sounds of different frequencies [6]. A lot of opinions concerning the frequency content of the heart sounds and murmurs have been reprented. The first heart sound has a slightly lower frequency than the cond heart sound [14]. The frequencies of the murmurs are higher [2], [15], and frequencies from 600 to 1000 Hz may exist [15].
Anand [6] stated that the frequency range of the heart sounds is up to 100 Hz. Arnott et al. concluded that the major concentration of energy for both the first and the cond heart sounds is below 150 Hz [16]. Cameron & Skofronick [17] and Rushmer [17] agreed with Arnott; both stated that the frequency range of the heart sound is from 20 to about 200 Hz. Vander et al. [12] suggested a bit higher frequency range, from 40 to 500 Hz.夜宵和宵夜的区别
The heart sounds and murmurs are measured from the chest wall. For the cardiac auscultation, four surface areas of the chest wall are suitable [5], [15]. The areas are called the aortic, pulmonary, tricuspid and mitral areas corresponding to the respective valvular areas [5], [15]. With the stethoscope placed in any one of the areas, the sounds from all the other valves can still be heard, though the sound from the special valve is the loudest [5].
III.M EASUREMENT SYSTEM
A.EMFi transducer
The transducer was implemented by using commercial EMFi material. Mechanical vibrations caud by the cardiac structure are converted into an electrical signal by the transducer. The operation is bad on the thickness variations of the EMFi material caud by the vibrations. The change in thickness generates a corresponding charge, and hence, a voltage to appear at the electrodes [11]. Thus the vibrations applied to the film can be detected by measuring either the charge or the voltage of the transducer. The transducer was chon to be a two-layer folded structure with a shape of a rectangle (size 7 cm x 5 cm). The film ud was metallized on both sides by sputtering to provide electrodes. The thickness of the electrodes is only a few dozens of nanometers, and thus the conductivity of the electrodes was ensured with a layer of aluminum foil. In the constructed transducer, the outer metal surface acts as a grounded electric shield and the inner surface of the two layer stack as a signal electrode. Finally, the outer surface of the EMFi transducer was strengthened with a layer of plastic sheeting. Connections were made by using a coaxial cable. The structure and operation principle of the EMFi transducer is described more precily in [18].
B.Measurement electronics
The electronics is enclod inside a plastic housing including amplifier electronics and a power source. Two 9 V alkaline batteries are ud as power source. The amplifier electronics is realized with a transimpedance amplifier followed by a voltage amplifier. AD820 (Analog Devices) amplifier is ud for both stages. The transimpedance amplifier converts the charge flow of the EMFi transducer to a voltage, which is further amplified with the voltage amplifier. The amplifier has the lower cut-off frequency (-3 dB) 1.5 Hz and the upper one 830 Hz.
C.Measurement tup
The EMFi transducer was attached to the chest wall by using double-sided tape. The transducer area corresponds to the pulmonary area, that is, the transducer is attached to the left side of the sternum, below the collar bone.
The measurement electronics was connected to the computer by using National Instruments SCB-68 connector block. The data was collected and analyzed with Matlab®software. The sampling rate f s = 1000 Hz was ud. Each measurement lasted for 30 conds.
D.Data analysis
In the data analysis, the Matlab® software was ud. The original raw data signal was filtered and the power spectral density (PSD) plot was calculated. In the filtration, the digital Butterworth cond order band pass filter was ud to remove the low-frequency respiration and heart pulsation components of the signal. The pass band of the filter was
chon to be from 20 Hz to 499 Hz, bad on the frequency content of the heart sounds discusd in Section 2. The heart sounds can be detected within this pass band. The 50 Hz noi was removed with an elliptic band stop filter. The PSD estimate was computed by using the Welch method of averaged periodograms [19].
IV. TEST MEASUREMENTS
For the test measurements, the EMFi transducer was taped to the chest wall of a healthy test person by using double-sided tape. In Fig. 1, the measured raw data signal is shown. The main component of the signal is respiration. The respiration frequency is eight times in 30 conds. Also the pulsation of the heart can be distinguished as sharp spikes.
猪蹄炖山药Time [s]
V o l t a g e [V ]
Fig. 1. Measured raw data signal. The main component of the signal corresponds to respiration.
In Fig. 2, the filtered signal is shown. To distinguish the parate heart sound components, only a ction of four conds of the signal is shown. In the figure, five cardiac cycles can be en, each consisting of the first heart sound (S1) and the cond heart sound (S2).
屡见不鲜
伟大的友谊
-3
Time [s]
V o l t a g e [V ]
Fig. 2. Filtered signal. Each cardiac cycle consists of the first (S1) and the cond (S2) heart sound.
PSD plot provides information of the frequency content of the heart sounds. The plot was computed from the filtered
30 conds signal by using 2048-point periodograms and 50% overlapping. In Fig. 3, the PSD plot of the filtered signal is shown.
10
10
10
10
10
-8
10
-6
Frequency [Hz]
P S D [V 2/H z ]
Fig. 3. Power spectral density (PSD) plot.
The raw data signal was also measured by holding breath. A ten conds ction of the measured signal is shown in Fig. 4. With this measurement tup, the main component is the pulsation of heart. Fourteen peaks are visible and thus the heart rate has been 84 during the measurement. Fig. 5 reprents a four conds ction of the filtered signal.
Time [s]
V o l t a g e [V ]
Fig. 4. Raw data signal measured by holding breath. The heart pulsation can be en as sharp spikes.
-3
Time [s]
V o l t a g e [V ]
Fig. 5. Filtered signal measured by holding breath. Five cardiac cycles consisting S1 and S2 are shown.
The frequency content of the signal is mainly the same as
it was in the previous ca. Therefore the PSD plot is not reprented here.
V.R ESULTS
The raw data signal measured with respiration reveals the respiration frequency. With proper filtering, the heart sound components can be en. The first and the cond heart sounds are clearly visible. The signal was measured from the pulmonary area, and thus the cond heart sound has occasionally larger amplitude. However, the amplitude of the heart sounds slightly alters due to the pha of the respiration. Bad on the power spectral density plot shown in Fig. 3, the frequency content of the heart sounds is concentrated on frequencies up to 80 Hz. This corresponds to the values discusd in Section 2.
The pulsation of the heart is clearly en from the raw data signal measured by holding breath. Again, in the filtered signal, the cond heart sound has slightly larger amplitude than the first sound due to the measurement area. Since the measurement was carried out by holding breath, the amplitudes of the heart sounds do not alter. With this measurement tup, the heart sounds are more evident. Some noi appeared in the filtered signals. The amount of noi could be reduced by
more efficient shielding of the EMFi transducer. However, the extraneous nois did not disturb the measurement. The EMFi material records the mechanical vibrations and is not as nsitive to extraneous nois coupled via air as normal microphones. Hence it is suitable material for measurements of this kind.
VI.C ONCLUSIONS
In this study, a new measurement system for the heart sounds was developed. The transducer was implemented by using ElectroMechanical Film (EMFi). The EMFi transducer converts the mechanical vibrations caud by the cardiac structure to an electrical signal. Thus the operation is more straightforward than in conventional stethoscopes, where a bell is required as an impedance matcher between the air and skin [17] before the conversion to an electrical signal is feasible. In addition, the amplifier electronics for the EMFi transducer was implemented. The measured and amplified signal was transmitted to the computer and analyzed with the Matlab® software.
With the constructed EMFi transducer and amplifier electronics, some test measurements were carried out. By taping the transducer to the chest wall of a healthy test person and by using proper digital filters, the heart sounds can be clearly en. In the test measurements, the signal was measur
ed both with respiration and by holding breath. With the latter one, the heart sound signal was more distinct, since the pha of the respiration did not affect the amplitudes of the heart sounds. From the signal measured by holding breath, the heart rate can be calculated. In addition, the respiration frequency can be determined from the signal measured with respiration.
The EMFi transducer taped to the chest wall reveals a lot of valuable information. The respiration frequency, heart rate and heart sounds can be determined, although the focus here was in the measurement of heart sounds. With EMFi material and by using more sophisticated manufacturing techniques, it would be possible to implement a plaster-like transducer measuring vital signals. Due to the low ba material cost, the transducer could be disposable. Bad on the results obtained here, the EMFi ems to be a promising material for physiological measurements.
武昌起义领导人R EFERENCES
扬州科技馆[1]T. Reed, N. Reed and P. Fritzson, “Heart sound analysis for symptom
detection and computer aided diagnosis”, Simulation Modelling
Practice and Theory, vol. 12, pp. 129-146, May 2004.
[2]Z. Zhidong, Z. Zhijin and C. Yuquan, “Time-frequency analysis of
heart sound bad on HHT”, Int. Conf. Communications, Circuits and Systems, 2005.
[3]L. T. Hall, J. L. Maple, J. Agzarian and D. Abbott, “Sensor system for
heart sound biomonitor”, Microelectronics Journal, vol. 31, pp. 583-
592, July 2000.
[4]J. Webster (ed.), “Medical instrumentation: application and design”,
Boston: Houghton Mifflin, 1978, ch. 7.
[5] A. Guyton, “Textbook of medical physiology”, Philadelphia: W.B.
Sounders, 1976, 5th edition, ch. 27.
[6]R. S. Anand, “PC bad monitoring of human heart sounds”,
Computers and Electrical Engineering, vol. 31, pp. 166-173, March
2005.
[7]S. R. Mesr, J. Agzarian and D. Abbott, “Optimal wavelet denoising
for phonocardiograms”, Microelectronics Journal, vol. 32, pp. 931-
941, December 2001.
[8]L. Torres-Pereira, P. Ruivo, C. Torres-Pereira and C. Couto, “A
noninvasive telemetric heart rate monitoring system bad on
phonocardiography”, IEEE Int. Symposium on Industrial Electronics,
1997, pp. 856-859.
[9] A. Zucherwar, R. Pretlow, J. Stoughton and D. Baker, “Development
of a piezopolymer pressure nsor for a portable fetal heart rate
monitor”, IEEE Trans. Biomedical Engineering, vol. 40, pp. 963-969,
September 1993.
[10]M. Paajanen, J. Lekkala and K. Kirjavainen, ”ElectroMechnical Film
(EMFi) – a new multipurpo electret material”, Sensors and
Actuators A: Physical, vol. 84, pp. 95-102, August 2000.
[11]Homepage of Emfit Ltd, , 2.4.2007.
[12] A. Vander, J. Sherman and D. Luciano, “Human physiology: the
mechanisms of body function”, New York: McGraw-Hill, 1990, 5th
edition, ch. 3C.
[13]S. M. Debbal and F. Bereksi-Reguig, “Time-frequency analysis of the
first and the cond heartbeat sounds”, Applied Mathematics and
Computation, vol. 184, pp. 1041-1052, January 2007.
[14] B. Kirby and K. MacLeod, “Clinical examination of the heart”,
Medicine, vol. 34, pp. 123-128, April 2006.
[15]R. Rushmer, “Cardiovascular dynamics”, Philadelphia: W.B.
Sounders, 1976, 4th edition, ch. 11.
[16]P. Arnott, G. Pfeiffer and M. Tavel, “Spectral analysis of heart sounds:
relationships between some physical characteristics and frequency
spectra of first and cond heart sounds in normals and
hypertensives”, Journal of Biomedical Engineering, vol. 6, pp. 121-
128, April 1984.
[17]J. Cameron and J. Skofronick, “Medical physics”, New York: John
Wiley & Sons, 1978, ch. 8.
[18]S. Kärki, T. Salpavaara and J. Lekkala,”EMFi in wearable audio
applications”, 4th Int. Workshop on Wearable and Implantable Body
Sensor Networks, 2007.
[19]P.D. Welch, “The u of fast Fourier transform for the estimation of
power spectra: a method bad on time averaging over short, modified periodograms”, IEEE Trans. Audio and Electroacoustics, vol. 15, pp.
70-73, 1967.
