RANDOM $58M). ULTRASGMC amon- FLOW @EFESTEON mesis fer the Degree 0? m. 3.. MECHW 3mm uwvzasm FRANK mm m _ 2973 ”@firgyfifljgflflflfiyfljflifiy “BRIE;- J Eiahigan ; w:3 University MTCHTGANL STATE UN? VERSTTY BRARY w ABSTRACT RANDOM SIGNAL ULTRASONIC BLOOD FLOW DETECTION BY Frank Saggio III Although conventional pulsed or continuous ultra- sonic Doppler systems are used for transcutaneous blood flow detection with considerable success, it is found that these systems are severely limited by signal-to-noise ratio problems; particularly when used on deep-lying vessels such as the superior vena cava, or on patients with emphy- sema where good signals are difficult to obtain.1 In addition, because of the periodic nature of these conventional signals, ambiguities in target range or blood velocity measurement may occur even when examining periphereal vessels. The objective of this thesis is to pr0pose and exam- ine concepts which will improve the performance of ultrasonic blood flow sensing systems. A random signal ultrasonic system is deScribed which uses a digital correlator and several bandpass filters to extract the target range and velocity information. A model of the random signal system is presented and analyzed to give the basic equations describing the system. Longitudinal Frank Saggio III range and velocity resolution are shown to be independently controllable. The performance capabilities of the random signal system are compared to those of a conventional pulsed system. An experimental random signal ultrasonic system used. to make in-vitro Doppler frequency measurements is described. The transmitted signal consists of 10 usec pulses of random noise with a center frequency of 4.59 MHz and a bandwidth of 4 MHz. Results are presented which indicate that the experimental system is capable of measuring the Doppler difference frequency within an error range of :_l%. The theory and operation of a dual element broadband transducer designed for ultrasonic blood flow detection is presented. The novel transducer is compared with a single element transducer in terms of lateral range resolution. The experimental results indicate a marked increase in lateral resolution with the dual element transducer. Two methods for determining the angles associated with Doppler blood flow measurement are described. 1K. L. Gould, D. J. Mozersky, D. E. Hokanson, D. W. Baker, J. W. Kennedy, D. S. Sumner, and D. E. Strandness, Jr., "A noninvasive technique for determining potency of saphenous vein coronary bypass grafts," Circu- lation, Volume XLVI, pp. 595-600, September 1972. RANDOM SIGNAL ULTRASONIC BLOOD FLOW DETECTION BY Frank Saggio III A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Electrical Engineering and Systems Science 1973 To all of my families .1 .i ACKNOWLEDGMENTS I thank Dr. C. P. Jethwa for his assistance and guidance throughout the duration of this project. The labor involved in writing this thesis was considerably lightened by his help in preparing a detailed outline. I wish to thank Philip A. Priest for exhibiting an occasional counterexample of excellence. He is responsi- ble for many of the drawings found in this thesis. Appreciation is extended to the General Motors Corporation for providing the financial support which has made my graduate study and this thesis possible. TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . vi LIST OF FIGURES . . . . . . . . . . . . vii Chapter I 0 INTRODUCTION 0 O O O O O O I O O O 1 1.1 The Importance of Blood Flow Studies . . l 1.2 Review of the Existing Blood Flow ' Meters . . . . . . . . . . . . 2 1.2.1 Transmission-type Ultrasonic Flow Meters . . . . . . 3 1.2.2 Reflection-type Ultrasonic Flow Meters . . . . . . . 6 1.3 Preview of the Random Signal Ultra- sonic System . . . . . . . . . . 12 II. THEORY AND GENERAL DESCRIPTION OF THE RANDOM SIGNAL ULTRASONIC SYSTEM . . . . . . . 14 2.1 Introduction . . . . . 14 2.2 Basic Model of the Random Signal Ultrasonic System . . . . . . . . 14 2.3 Theory of Operation . . . . . 17 2.4 Characteristics and Advantages of the Random Signal System . . . . . . 24 2.5 Limitations of the Random Signal System . 27 2.6 General Description of the System . . . 28 III. EXPERIMENTAL RANDOM SIGNAL ULTRASONIC SYSTEM . 35 3.1 System Implementation . . . . . . . 35 3.1.1 Circulation Model . . . . . . 35 3.1.2 Transmitter . . . . . . . . 38 3.1.3 Transducers . . . . . . 39 3.1.4 Adjustable Delay Line . . . . 40 3.1.5 Receivers . . . . . . . . 43 3.1.6 Correlator . . . . . . . . 43 3.2 Do oppler Resolution Results . . . . ._ 46 3.3 Conclusions . . . . . . . . . . ‘ 53 iv Chapter Page IV. THEORY AND OPERATION OF A DUAL ELEMENT BROADBAND TRANSDUCER . . . . . . . . . 54 4.1 Introduction . . . . . . . . . . 54 4.2 Principle of Operation . . . . . 55 4. 3 Experimental Results and Conclusions . . 57 V. METHODS FOR DETERMINING THE ANGLES ASSOCIATED WITH DOPPLER BLOOD FLOW MEASUREMENTS . . . 68 5.1 Introduction . . 68 5. 2 Oblique Triangulation Method 69 5. 3 Compound B-scan Technique . 5.3.1 Introduction . . . 70 71 5.3.2 I 5.3.3 Calculation of the Angle of Attack 75 5.3.4 . Basic Principles . . 72 Conclusions . . . . . . . 81 VI. SUMMARY . . . . . . . . . . . . . 82 6.1 Results and Conclusions . . . . - 82 6. 2 Recommendations for Further WOrk . . . 84 REFERENCES . . . . . . . . . . . . . . 86 A} \[{([.[l’l\{{ (iii [4! II IIII‘III.’ Ill LIST OF TABLES TABLE Page 1. Characteristics of the Experimental System . . 47 2. Comparison of Theoretical D0ppler Frequencies Expected and Experimental Doppler Frequencies Observed . . . . . . . . . . . . 52 vi Figure 1. 13. 14. 15. 16. LIST OF FIGURES Arrangement of Transducers for Measurement of Flow Velocity by Transit Time Difference in Up and Down Stream Paths . . . . Arrangement of Transducers for Measurement of Flow Velocity by Reflection Methods Basic Block Diagram of the Random Signal Ultrasonic System . . . . . . . Expected Correlator Output . . . . . Diagram of the Complete Random Signal Ultrasonic System .° . . . . . . Waveforms Associated with Blood Velocity Measurements . . . . . . . . . Simplified Experimental System . . . Circulation System . . . . . . . Transducer Arrangement for Doppler Detection Adjustable Analog Delay Line . . . . Photograph of the Experimental System . Frequency Domain Representation of Background Noise . . . . . . . . . . . Doppler Resolution . . . . . . . Doppler Resolution . . . . . . . Doppler Resolution . . . . . . . Construction of the Dual Element Wideband Ultrasonic Transducer . . . . . . vii Page 15 20 29 33 36 37 41 42 45 48 49 50 51 56 Figure Page 17. Echoes Received by a Single Element Trans- ducer Scanning the Target Laterally . . . 60 18. Echoes Received by a Single Element Trans- ducer Scanning the Target Laterally . . . 62 19. Echoes Received by the Dual Element Trans- - ducer Scanning the Target Laterally . . . 64 20. Echoes Received by the Dual Element Trans- ducer Scanning the Target Laterally . . . 66 21. Comparison of the Lateral Resolution of Single Element and Dual Element Transducers . . . 67 22. Oblique Triangle Construction . . . . . . 71 23. A-scope and B-sc0pe Displays of Ultrasonic Pulse-Echo Information . . . . . . . 74 24. Several Lateral Compound B-scans . . . . _. _ 76 25. Longitudinal Compound B-scan . . . . . . 77 26. Determining the Angle a . . . . . . . . 79 27. Determining the Angle 8 . . . . . . . . 80 viii CHAPTER I INTRODUCTION 1.1 The Importance of Blood Flow Studies Volume blood flow measurement is important in deter- .mining the amount of oxygen being delivered to various re- gions of the body. Also, knowledge of the acceleration of the blood flow in the aScending aorta or other large vessels may be useful in assessing the performance of the heart [5]. This volume blood flow information is normally obtained by measuring the mean blood flow velocity and the vessells cross-sectional area at a particular site. The determination of a velocity profile in arteries may be useful in identifying arteriosclerosis by distinguish- ing a laminar flow from a turbulent flow. This method may also provide a basis for the early identification of turbu- lent areas in atherosclerotic arteries. Blood velocity profile, volume flow and average flow velocity measurementsalso provide information about the blood viscosity, which is known to be an important factor in the etiology of ischemic diseases [24]. From the above discussion it is clear that quantita- tive blood flow information is often essential before a particular disease can be properly diagnosed. At present, cardiac catherization and angiography techniques are used to access many of the aforementioned physiological para- meters. However, these invasive techniques are limited in their routine use in postoperative studies because of the potential mortality associated with their repeated applica- tion [6,26]. For this reason there is a strong interest in the development of a noninvasive ultrasonic method for measuring these parameters. This thesis describes the feasibility of such an ultrasonic system which can measure with a high degree of spatial resolution, many of the above physiological quantities. The following section summarizes the conventional ultrasonic methods and devices for determining the blood flow velocity. 1.2 Review of the Existing Blood Flow Meters Several different techniques, from plethysmography to nuclear magnetic resonance, have been used for detection and measurement of blood flow. The major difficulty with these methods is the lack of specificity required to sense flow at a particular site [5]. Ultrasonic techniques do not suffer from such handicaps and hence are ideally suited for medical research, diagnosis and treatment. Franklin [30], Wells [22], and Baker [5] have pub- lished excellent surveys of various ultrasonic blood flow meters. The existing ultrasonic devices can be classified into one of two broad categories: the transmission type or the reflection type. Continuous or pulsed ultrasonic waves can be used with either type flowmeter. 1.2.1 Transmission-type UltrasonicTFIow Meters These flow meters are based on the principal that a sound beam, passed diagonally through a blood vessel, would exhibit a difference in the time required to traverse the vessel alternately in the upstream and downstream di- rections. The average blood velocity v_of the medium can be determined by using the following relationship: v =' m; - c2 (1) ‘2d . cose where At is the measured difference between the up- stream and downstream transit times, c is the velocity of ultrasound relative to the propagating medium, d is the distance between two transducers placed diagonally on either side of the vessel, and 6 is the angle between the axis of ultrasound propagation and the flow axis. Figure 1 illus- trates a typical transducer arrangement. Many transmission type flow meters have been de- signed. Kalmus [31] developed an instrument to measure fluid velocity by noting that a continuous ultrasonic wave transmitted alternately upstream and downstream would produce a phase difference proportional to the average velocity of the medium through which the sound passes. Haugen gE_gl. [32] and Baldes gt_al. [33] used modified versions of the instrument developed by Kalmus. Franklin gt_al. [34,35] experimented with a pulsed ultrasonic flow meter. The upstream and downstream transit times for 3 MHz ultrasonic pulses were compared by consid- ering the amplitudes of corresponding ramps generated by a time-to-voltage analog converter. Farall [36] improved the earlier system of Baldes [33] by using two transmitter and receiver pairs. The reference signal from each transmitter was compared with its respective received signal by a phasemeter. The difference between the outputs of the two phasemeters was proportional to the flow velocity. Zarnstroff gt_§1. [37] reported an instrument in which a continuous 1 MHz signal was transmitted alternately upstream and downstream. The received signal was heterodYned to 10 KHz. Differences in the upstream and downstream phase of the 10 KHz signal were proportional to the flow velocity. Noble gt_gl, [38] reported a phase shift technique which eliminated the need for switching the sound path upstream and downstream, and hence improved the high frequency response of the system. The performance of transmission type flowmeters is very unsatisfactory at velocities below 10 cm/sec. But the ,//r———-TRANSDUCER ‘ {'— VESSEL .//<\\\ TRANSDUCER Figure'1.—qArrangement of Transducers for Measurement of Flow Velocity by Transit Time Difference in Up and Down Stream Paths. main disadvantage associated with these flow detectors is that surgery must be performed to fasten the transducers and holder directly to the blood vessel. Consequently as reported by Wetterer [39], the transmission type ultrasonic flow meter is no better than the electromagnetic flow de- tector. .2.2 Reflection-type ltrasonic Flow Meters l U These flowmeters are based on the principal of the Doppler effect. When ultrasound is transmitted into the blood stream, a portion of the signal is scattered by the moving blood particles. The frequency of the back-scattered sound is Doppler shifted. The relationship between the average Doppler difference frequency fd and the average blood flow velocity v is given by the following equation [22,40] V=C°fd (2) E;7733§E — cosBT where c is the velocity of ultrasound in blood, fo is the transmitted ultrasonic frequency, a is the angle between the transmitted ultrasonic beam and the velocity vector, and B is the angle between the reflected ultrasonic beam and blood velocity vector. Figure 2 illustrates a typical transducer arrangement. l<: TRANSDUCERS' Figure 2.-qArrangement of Transducers for Measurement of Flow Velocity by Reflection MethOds. Reflection-type devices may Operate in the con- tinuous wave mode or the more sophisticated pulsed Doppler mode. Satomura [42] was first to demonstrate continuous wave Doppler motion detection in 1956. Franklin gt_gl. [43] used this technique for blood velocity detection in animals. Baker [44] developed a practical instrument for the trans- cutaneous detection of blood flow in humans. The early continuous wave flow meters were not capa- ble of detecting the direction of the blood velocity vector with respect to the transducer. In 1966 McLeod [45,46] used a version of quadrature phase detection common to single sideband radio techniques to demonstrate the first success- _ful direction sensing Doppler device. Yoshitoshi gt_gl. [47] described another method of direction sensing. Continuous wave blood flow detectors have found wide application in medical research. However, one of the greatest shortcomings of the continuous wave device is its inability to obtain a measurement of range. The continuous wave technique is inherently sensitive to all movement along the length of the sound beam. This lack of range sensitivity prevents the measurements of vessel diameter and velocity profile. In order to preserve the range information, Moving Target Indicator (MTI) or pulsed DOppler radar techniques have been employed. Barnes and Thurstone [49] have devel- oped an MTI ultrasonic system which uses a delay line . canceler to filter the clutter echoes and pass only the DOppler frequency shifted signals returned from moving targets. However, this system is limited by the additional complexity required to achieve special filter characteristics which maximize the output signal-to-noise ratio. The tech- nique is also hampered by the need to maintain perfect delay line adjustments in spite of temperature changes. For these reasons the pulse-Doppler system is preferred to the MTI system. In the pulse-Doppler system a burst of ultrasound is transmitted. The distance to the dynamic target and its velocity with respect to the ultrasonic beam are determined by range-gating the return echoes. These received echoes are both phase and amplitude modulated by the moving target. The Doppler frequency or velocity information can be ob- tained by non-coherent or coherent detection. The non-coherent pulsed-Doppler system makes use of the amplitude fluctuations in the echo signal to recog- nize the Doppler component produced by a moving target. Range-gating is used to select an echo for detection from a pre-determined depth. A linear or logarithmic receiver is necessary to preserve the amplitude components. The Doppler difference frequency is detected from the range- gated sample by a square law diode detector. The primary difficulty with the non-coherent technique is the erratic 10 fluctuation in the detected Dopplersignal due to the pres- ence of noise clutter components. No blood flow detectors based on this principal are reported in the literature. In a coherent pulsed-Doppler system, the range- gated echo is compared with a sample of the original trans- mitted signal in a phase detector to extract the Doppler components. A number of coherent systems have been built and used for measuring blood velocity, volume blood flow, and for mapping the lumen of the blood vessels [5,50-54]. Conventional pulsed-Doppler systems are generally inadequate for transcutaneous blood flow measurements in deep-lying vessels such as the ascending aorta, superior vena cava, or renal artery [6,23]. A significant improve- ment of the system's signal—to—noise ratio is required to increase the detection depth without exceeding the safe ultrasonic intensity level established for humans [27,28]. An increase in lateral resolution is also required to apply the pulsed-Doppler methods to very small vessels. It is known that if n pulses each with the same signal-to-noise ratio were ideally integrated, the resultant signal-to-noise ratio would be exactly n times that of a single pulse [55]. However conventional pulsed-Doppler systems cannot use this method of enhancing signal-to- noise ratios without destroying the range information. The range resolution capability of conventional pulse- echo systems is normally increased by narrowing the 11 transmitted pulse. Two basic difficulties arise with this method. The first is generating and processing the transmitted and received signals. The second is maintain- ing sufficient average power in the transmitted signals. The usual requirement for a non-ambiguous range determina- tion limits the maximum pulse repetition frequency. Thus, as the pulse width is narrowed, the amount of peak power re- quired to maintain sufficient average power soon becomes unreasonable. _Perhaps it may exceed the safe intensity level for tissues. The solution to these inherent problems associated with conventional pulse-Doppler lies in the use of trans- mitted signals having a larger time-bandwidth product than a simple pulse [62]. The transmitted waveform may have a longer duration for the same range resolution capability, as in pulse compression or pseudo-random phase coded tech- niques. Or the signal may be transmitted at a higher repetition rate without range ambiguities, as in pseudo- noise sequences or signals using pulse-to-pulse frequency hopping. Wagg and Gramiak [56,57] have developed a system utilizing the continuous transmission of a carrier which is phase modulated with a pseudo-random binary sequence. They have reported an enhancement in the signal processing gain over the conventional pulse-Doppler system. However, 12 the improvement which is obtained with this technique is somewhat limited by the complexity of the equipment required to generate and process these signals. Also, because of the periodic nature of these signals, ambiguities in target range or velocity may still occur. . 'l;3.Preview of the Random Signal Ultrasonic System A random signal pulse-Doppler system.may provide the solution to the above problems. The use of transmitted sig- nals derived from wideband Gaussian noise permits extremely large time-bandwidth products to be obtained and eliminates the difficulties with range and velocity ambiguities charac- teristic of periodic signals. Range and velocity resolution can be shown to be independently controllable. The signal- to-noise ratio may be enhanced many times by integrating the returns from successive pulses without destroying spatial resolution. In Chapter II a generalized model of the random signal pulse-Doppler system is presented and analyzed to give the basic equations describing the operation of the system. Chapter III describes the experimental system built and the results obtained for the Doppler resolution. Chapter IV discusses the theory and operation of a novel transducer which improves the lateral resolution of the pulse-Doppler system. 13 In Chapter V methods for determining the angles associated with ultrasonic blood flow measurement are described and analyzed. Chapter VI presents a summary of the results obtained and some recommendations for further experimental work. CHAPTER II THEORY AND GENERAL DESCRIPTION OF THE RANDOM SIGNAL ULTRASONIC SYSTEM 2.1 Introduction The theory and effectiveness of the random signal microwave system are well documented [58-61]. In order to apply this technology to biomedical measurements it is nec- essary to change the transmitted signal from microwave electromagnetic radiation to wideband acoustical energy which is more suitable for propagation in living tissues. The concepts of resolution, accuracy, ambiguity, detection and clutter which are common in radar discussion, have their counterparts in the ultrasonic system with appropriate changes in the scale factors. A brief discussion of the theory of operation and a general description of the random signal ultrasonic system are presented in the following sections.‘ 2.2 Basic Model of the Random Signal Ultrasonic System The basic block diagram of the random signal system is shown in Figure 3. The transmitted signal x(t) is as- sumed to be a sample function from a stationary zero mean 14 15 usmuso .Eoumhm owsommuuab Hogwm Bong on» no Emu—manna “~0on “sensual...” cashew H033 anagrams . Aqu IIII'IIII naumaouuoo fits + C. I use + H: x . Com n A»; 3K van—Boom? A»: Hogan pmbwmomm my - 2.5 + or. u E... Amanda monoummom u a wuss Swamp as >306 E OUHDOW 37 "new? "momma Eats?" Erasmus: ./ .3230?” uoflgmsdsu 16 random process generated by a random noise source. A por- tion of the transmitted signal is used to simulate the de- lay and Doppler shift of a fictitious target. This signal r(t) becomes the reference input to the correlator. The rest of the transmitted electrical signal is converted to acoustic energy by applying it to the input of a piezoelec- tric transducer which is used as'a transmitter and is aimed at the living tissue. The reflected acoustic signal is converted back to an electrical signal by a piezoelectric receiving transducer. A single transducer is normally ade- quate for both transmission and reception of pulsed signals. The reflected signal y(t) will be an amplitude scaled and delayed version of the transmitted signal plus it will contain a noise term. This signal provides the other in- put to the wideband correlator. The output of the corre- lator z(t) is the estimate of the cross-correlation function of y(t) and r(t). By providing a bandpass filter after correlation the effects of the noise term are reduced. Inaddition, those portions of the cross-correlation function which do not rep- resent targets of interest are eliminated. Thus the band- pass filter provides both the correlator smoothing time and the desired velocity resolution. This point will be dis- cussed in detail later on. Although the actual implementation of the cross- correlator is digital, it is simpler to discuss the theory 17 in terms of an equivalent analog system. The digital processor achieves the same end results; the only basic difference is the requirement of longer processing time in the digital version. Unless otherwise stated, the analysis of Chapter II will assume the ultrasonic system is perform- ing a continuous cross-correlation. 2.3 Theory of Operation For ease in discussion let the transmitted signal x(t) be a sample function from a stationary random process having a spectral density that is symmetrical about some center frequency fc. Suppose at time t = 0, there exists a single point target at range r, moving with a constant radial velocity \n. Then the round trip time delay I and the delay rate a associated with the signal returned from this target are given by {=23 . (3) C -2v .a: “at; c ‘4’ where c is the propagation velocity of the acoustic signal in living tissues. The returned signal y(t), consisting of the reflected signal from the single point target and the system noise, can therefore be written as y(t) = am -x[(1 + out - 11+ n ' <5) 18 where the attenuation coefficient a(r) is real and between zero and one, and n(t) is the system noise term. The noise term n(t) is assumed to be zero mean and independent of x(t). This received signal y(t) is cross-correlated with a reference signal r(t) which has the following mathematical form r(t) = x[(1 + ar]t - If] V (6) with Tr and or the delay time and delay rate respectively for the reference signal r(t). These parameters (Tr and ar) can be adjusted at will and can be made to match a particular target reflecting point. V v The correlator output z(t) provides an estimate of the cross-correlation function. The expected value of z(t), which is the desired correlator output, is given by E{2(t)} = E {[a(I)x{(l+a)t-T] + n(t)} ° x[ (1+ar)tfrr>a}» ‘ (7) or, since n(t) is assumed to be zero mean and independent of x(t), E{z(t)} = aTIT'- th(a r-a)t-TTr-T)]- (8) where Rx (-) is the autocorrelation function of x(t), and air] is the mean value of the attenuation coefficient. 19 Since Rx(-) is the autocorrelation function of a bandpass random process, it can be written in terms of a slowly vary- ing envelOpe and cosine term E[z(t)] = a??? ' Rc[(q r-a)t-(Ir-r)]coswC - [(Ot. r-Mt-(Tr-TH (9) where RC(-) is the envelope of the correlation function and we is the center frequency about which the spectral density of the transmitted signal x(t) is symmetrical. The expected correlator output (equation 9) is de- picted in Figure 4. The radian frequency of the output is w = NC ' (Gr - a) (10) and the time at which the peak occurs is Tr, T t = —-———- (11) Figure 4 is a replica of the autocorrelation function of x(t) except that it has been stretched in time by a scale factor of 1/(ar - a). Hence it will have a center frequency of (ar - a) ° fc and this frequency will typically be in the audio range. Therefore a suitable filter at the output of the correlator will accept or reject this component as de- sired (See Figure 3). Since a depends upon the velocity of the reflecting point, points with different velocities will produce outputs with different center frequencies. Thus 20 Figure 4.—«Expe¢ted Correlator Output. 21 the ability of the ultrasonic system to resolve targets at the same range but having different velocities depends-pri- marily upon the bandwidth of the filter at the output of the correlator. If the bandwidth of this filter is W, then by combining equations 4 and 10 it is easy to show that the velocity resolution of the system will be on the order of - C where c is the propagation velocity of the acoustic wave in living tissues (approximately 1.48 x 105cm/sec). For a bandlimited transmitted signal with center frequency fc = 5 MHz and the bandwidth of the filter following correlation equal to W = lOHz, the velocity resolution Av is about .15 cm/sec. When reflecting points are separated in range their responses in the correlator output will occur at different times. Hence, the ability to resolve targets having the same velocity but different target ranges depends upon the time separation of the responses (which have also been in- creased by the factor l/(cr - c))compared to the expanded width of the autocorrelation function. Again it is easy to show that if the transmitted signal has a bandwidth B, which determines the width of the correlation function, the range resolution AR is on the order of c , (13) AR 22 wherez> Alfeleevl AllDO‘IBA AIIOO‘SBA '< '< 2: I ms: AW: w s x i i: A m m E 3 ms: . e .A cums 08 8m: 8 o I T I _... V 04 a. €— 34 Once orientation has been achieved, the investigator can switch to the second mode of display shown in Figure 6d. Here velocity is depicted as a function of time at a specific range. _The entire system may also be synchronized with an electrocardiogram to provide additional information. CHAPTER III EXPERIMENTAL RANDOM SIGNAL ULTRASONIC SYSTEM 3.1 System Implementation A simplified version of the random signal ultra- sonic system described in Chapter II has been built and results have been obtained for the Doppler resolution. The block diagram of the experimental system is shown in Figure 7. 3.1.1 Circulation Model Figure 8 is a photograph of the circulation system used to simulate blood flow. Polystyrene particles of size 50-75 microns are mixed in water and circulated in the tygon tube by using a peristaltic pump. The flexible tygon tube is connected to a section of glass tubing which passes through the water tank. This glass tubing corresponds to a blood vessel with an inside diameter of 8mm and a wall thickness of 0.5mm. The transducers used for transmitting the ultrasonic pulses and detecting the Doppler frequency shifted echoes are positioned directly over the glass tube. Each transducer 35 T | water tank I l V~WW ——-> . . | bandpass -_ T filter V T S/ ;/ { T T transmittet 3 receiver , .pu Lps "’5 ~————|>———J -——-—~mI>*———“‘- i ) ‘ [___j 4— power splitter : gate ' Y . ; i random signal F——fi—~a r~«———L—————— noise enerator . c\ 3 g g ‘L\\ ' pre- . ‘ :>transducers Iamplifier ad ustable - ’ ' LI water | l i 'L amplifierj T | T T V delay line L~Wr——‘ T '—"’ ~_TT .—.~.. »_ ._._.-____..— amplifier. ~~ -1...“ ‘7 J digital L—._-—_*._- 1 »—— . l correlator T reference ‘ received signal “—‘*”“r"—‘“J echo signal . bandpass l filter t ““auai. amplifier ! r __ -- -,...- ._______-_J Y .“_I1_______fi Spectrum analyzer .- __._._._.-'_ _ _ -_.___J T Figure 7.-—Simplified Experimental System. 37 Figure 8.-'Circu1ation System. 38 face is separated from the centerline of the tube by a nominal distance of 3 centimeters. These piezoelectric transducers will be discussed in section 3.1.3. For the given tube diameter (8mm ID), the variable speed peristaltic pump is capable of moving the polystyrene particles over a continuous velocity range from 0-40 cm/sec. These particles circulate best at velocities above 10 cm/sec, due to their large size and corresponding weight. As shown in Figure 8 the bases of both peristaltic pump and polyethylene water tank are insulated with foam cushions to minimize the environmental vibrational effects. 3.1.2 Transmitter The transmitter subsystem consists of the following: 1. A random signal noise generator 2. A bandpass filter 3. A transmission gate 4. A linear power amplifier and power splitter. The random signal noise generator uses a gas dis- charge tube as its noise source. The noise output from the gas tube is amplified in a two stage amplifier before being input to the active bandpass filter. The two-pole bandpass filter has a center frequency fc:=‘4'59 MHz and a bandwidth of 4 MHz. An operational ampli- fier is inserted after the bandpass section to provide an additional 7dB signal gain. 39 The filtered noise is limited to pulses lOusec wide by means of a linear transmission gate. The pulse repetition frequency (PRF) is approximately equal to 10 KHz. The gate has been fixed at 10usec although other durations are possible. A broadband solid state power amplifier covering the frequency range of 250 KHz to 110 MHz is used to amplify the gated signal. A nominal gain of 50dB raises the random noise pulse to approximately 120 volts peak-to-peak. A BNC tee connector acts as the power splitter shown in Figure 7. The amplifier output is used to drive the two piezoelectric transmitting transducers. The need for two transmitting transducers will be discussed in the subse- quent sections. 3.1.3 Transducers The experimental system uses single element broad- band transducers. These piezoelectric transducers have a low Q which makes them suitable for pulse-echo operation. Separate transducers are used for transmitting and receiving the random signals to avoid the problem of decoupling the transmitter from the receiver. Referring back to Figure 7, there are two transmitter and receiver transducer pairs. One set of transducers is used to generate the delay between the reference signal and the received signal echoes. This arrangement will be 40 discussed in the next section. The other transducer pair is used to detect the D0ppler signals. This latter arrange- ment is shown in Figure 9. The transmitting and receiving transducers used for sensing the motion of the polystyrene particles are positioned directly over the simulated blood vessel. The transducer faces are immersed in a water bath to provide the acoustic coupling necessary for the transmission and reception of ul- trasound. Each transducer face is separated from the center- line of the vessel by a nominal distance of 3 centimeters. The angles a and 8 associated with the calculation of the fluid velocity (see equation 2) are measured directly with a protractor. 3.1.4 Adjustable Delay Line The necessary delay between the reference and re- ceived signals is obtained by an adjustable analog delay line, rather than by sampling and using digital shift reg- isters. The analog delay line, consisting of a water tank and two piezoelectric transducers, is shown in Figure 10. The variable delay of the reference signal is achieved by keeping the transmitting transducer fixed and moving the receiving reference transducer. For a velocity of sound in water of approximately 1.48 x 105cm/sec, the range factor is 13usec of delay per 41 Figure 9.-—Transducer Arrangement For Doppler Detection. 42 Figure 10.- 76 LATERAL LINES 0F SCAN - (b) Figure 24.v—Several Lateral Compound B-scans. 77 ZXLONGITUDINAL (a) SCAN LINE _>’ X ) X ()1 (b) BLOOD VESSEL Figure 25.——Longitudinal Compound B—scan. 78 compound scan is performed along this constructed line. The ' blood vessel should be displayed from this scan as depicted in Figure 25b. The angle a of the blood vessel relative to the sur- face of the skin can be computed from the photographic display depicted in Figure 25b as follows: 1. A straight line is drawn through the photographic blood vessel image. 2. The slope s of this line relative to the surface is of the skin at the point x2 (22) a: II 21% w where dy is the vertical displacement of the blood vessel, and dx is the horizontal increment. 3. a is then given by a = tan-ls (23) Figure 26 summarizes the above procedure. With this preliminary calculation completed, ultra- sonic blood flow detection may begin. The transducer posi— tioned by its theodolite holder is acoustically coupled to the skin at the point x2 (Figure 27). The angle B of beam incidence relative to the skin surface is read directly from the theodolite holder. With the aid of Figu'e 27, 6 can be easily computed. 79 BLOOD VESSEL -——\ Figure 26.——Determining the Angle a. 81 y = 130° - a - e (24). e = 180° - y (25) = a + B As a final check, the transducer Operating in the pulsed mode should be moved such that 6 = a + B a 90°. The Doppler frequency shift should be equal to zero. 5.3.4 Conclusions This technique determines the location of the blood vessel and computes 6 the angle of ultrasonic attack. It appears to be a theoretically sound approach to the problem if the observer is experienced with the interpretation of compound B-scans. To the author's knowledge, no such method of angular calculation has been presented in the literature to date. The main disadvantage associated with this technique is that compound B-scan storage devices are very expensive. The final accuracy of the angle measurement will have to determine if such an investment is worthwhile. CHAPTER VI SUMMARY ' 6.1 Results and Conclusions The analysis presented in Chapter II shows that the range resolution of the random signal ultrasonic system is determined by the bandwidth of the transmitted signal. The velocity resolution is determined by the bandwidth of the filter following correlation. These bandwidths are inde- pendently controllable, thereby making it possible to make very accurate range and velocity measurements simultaneously. The primary advantage of the random signal system is its freedom from range ambiguities. This permits the use of high pulse repetition rates to measure any given Doppler shift unambiguously in the absence of clutter. Spectrum folding, which would occur for a conventional ultra- sonic system used to detect high velocity targets, is com- pletely avoided. In comparison to conventional pulse systems, the duty factor of the transmitted random waveform can be kept large thereby providing a greater average power for a given peak .power. Increased sampling rates, duty factors and coherent or noncoherent integration all provide an improvement in the system's signal-to-noise ratio. 82 83 In Chapter III the experimental random signal ultra-- sonic system is described and the Doppler resolution results are presented. These results indicate that the experimental system is capable of measuring the D0ppler difference fre- quency within an error range of i 1%. This system does not by any means represent the ultimate capabilities of random signal ultrasonic techniques. It was built to demonstrate and verify the theoretical results. The basic system con- figuration and performance will continue to be improved in many ways since the initial construction. Previous experimental work by Newhouse gt_gl. [64] established the range resolution capabilities of the random signal ultrasonic system. ‘ * The theory and Operation of a dual element broadband transducer [67] designed for ultrasonic blood flow detection is presented in Chapter IV. The novel transducer is compared with a single element transducer in terms of lateral range resolution. The experimental results presented in that chapter indicate a marked increase in lateral resolution with the dual element transducer. In Chapter V, two methods for determining the angles associated with Doppler blood flow measurement are described. The first technique appears to be the accepted method for calculating these angles. The second method, developed by the author, makes use of the information obtained from several compound B-scans to determine the slope of the blood vessel 84 and the angle of attack. Experience is required in inter- preting the compound B-scans. Although the latter method is shown to be theoretically feasible, no experimental re- sults are presented on this tOpic. 6.2 Recommendations for Further work To fully demonstrate the superiority of the random signal ultrasonic system in comparison to conventional pulse Doppler techniques, experimental results indicating the random signal-to-noise ratio are required. Further experiments involving the use of smaller polystyrene particles (7-10 microns) or human blood should be conducted to determine the capability of the random signal system in detecting Doppler shifts from smaller sound scatterers. It is hoPed that by combining the unique lateral resolution property of the dual element transducer with the longitudinal resolution capability provided by random signal processing that velocity profiles can be obtained for deep-lying and small blood vessels. The experimental feasibility of employing the com- pound B-scan technique to the transcutaneous calculation of the ultrasonic angles of attack needs to be demonstrated.2 Finally some theoretical work must be undertaken to determine if the equation used to relate the velocity of the moving target to the Doppler shift frequency (equations 2 85 and 19) is still valid for bandlimited random processes. The derivation of this formula assumes that only a single frequency fo is transmitted. In the random signal ultra- sonic system, fo is interpretted as the center frequency of the bandlimited random signal generated. REFERENCES 86 REFERENCES 1. Franklin, D. L., Schegel, W. A. and watson, N. W., "Ultrasonic doppler shift blood flowmeter," Biomed. Seis, 2. Kaneko, Z., Shirashi, J., Omizo, H., Kato, K., Motomiya, M., Izumi, T. and Okumura, T., “Analyzing blood flow with sonograph," Ultrasonic 4, p. 22-23, 1966. ' 3. Strandness, D. E., McCutcheson, E. P., and Rush- mer, R. F., "Application of a transcutaneous doppler flow- meter in evaluation of occlusive arterial disease," Surg. Gynec. Obstet., 122, p. 1039-1045, 1966. 4. Strandness, D. E., Schultz, R. D., Summer, D. S., and Rushmer, R. F., "Ultrasonic flow detection. A useful technique in the evaluation of peripherial vascular disease," Am. J. Surg., 113, p. 311-320, 1967. * 5. Baker, D. W., "Pulsed doppler blood-flow sensing," IEEE Transactions on Sonics and Ultrasonics, Vol. SU-l7, No. 3, p. 170-185, July 1970. 6. Gould, K. L., Mozersky, D. J., Hokanson, D. E., Baker, D. W., Kennedy, J. W., Sumner, D. S., and Strandness, D. E., Jr., "A noninvasive technic for determining patency of saphenous vein coronary bypass grafts," Circulation, Volume XLVI, p. 595-600, September 1972. 7. Yoshida, T., Mori, M., Minura, Y., Kikita, G., Takagish, 8., Nakanishi, K., and Satomura, 8., "Analysis of heart motion with ultrasound doppler method and its clinical application," American Heart Journal 61, p. 61-75, 1961. 8. Yoshitoshi, Y., Machii, K., Seldiguchi, H., Mishina, Y., Ohta, S., Hanaoka, Y., Shimizu, S., and Kuno, H., "Doppler measurement of mitral valve and ventrical wall velocities," Ultrasonics 4, p. 27-28, 1966. 9. Feigenbaum, H., "Diagnostic ultrasound as an aid to the management of patients with paricardial effusion," 87 88 10. McDonald, I. G., Feigenbaum, H., and Chang, Sonia, "Analysis of left ventricular wale motion by reflected ultrasound. Application to assessment of myocardial function," Circulation, Volume XLVI, p. 14-25, July 1972. ll. Sweatman, T., Selzer, A., Kamagaki, M., and Cohn, K., "Echocardiographic diagnosis of mitral regurgita- tion due to ruptured chordae tendineae," Circulation, Volume XLVI, p. 580-586, September 1972. 12. Meyer, R. A. and Kaplan, S. K., "Echocardiogra- phy in the diagnosis of hypoplasia of the left or right ven- tricles in the neonate," Circulation, Volume XLVI, p. 55-64, July 1972. 13. Lundstrom, N. R., "Echocardiography in the diag- nosis of congenial mitral stenosis and in evaluation of the results of mitral valvotomy," Circulation, Volume XLVI,.p. 44-54, July 1972. 14. Tajik, A. J., Gau, G. T., Ritter, D. G. and Schattenberg, T. T., "Echocardiographic pattern of right ventricular diastolic volume overload in children," Circu- lation, Volume XLVI, p. 36-43, July 1972. 15. Fortum, N. J., Hood, W. P., Jr., and Craige, E., "Evaluation of left ventricular function by echocardio- graphy," Circulation, Volume XLVI, p. 26-35, July 1972. 16. Freeman, J. J., Hitt, J. S., Rinalds, J. A., "A system for the detection of mitral valvular disease," Proc. IEEE Kelly communication conference, 1970. 17. Fry, F. J., Heimburger, R. F., Gibbons, L. V., and Eggleton, R. C., "Ultrasound for visualization and modi- fication of brain tissue," IEEE Transaction on Sonic and Ultrasonics, Vol. SU-l7, No. 3, p. 165-170, July 1970. 18. Callagan, D. A., Rowland, T. C., and Goldman, D. E., "Ultrasonic doppler observation of the fetal heart," Obst. Gync., N. Y. 23, 637, 1964. 19. Fielder, F. D. and Pocok, P., "Foetal Blood - detector," Ultrasonics, 6, p. 240-241, 1968. » 20. Johnson, W. L., Stegall, H. F., Leim, J. N. and Rushmer, R. F., "Detection of fetal life in early preg- nancy with ultrasonic dOppler flowmeter," Obst. Gyn. N. Y. .26, p. 305-307, 1965. 89 21. Rushmer, R. F., Baker, D. W., and Stegall,. H. F., "Transcutaneous doppler flow detection as nondestruc- tive technique," J. Appl. Physiol. 21, p. 554-566,1966. 22. Wells, P. N. T., "Physical principles of ultra- sonic diagnosis," Academic Press, London, 1969. 23. Lathi, B. P., "Communication systems," John Wiley and Sons, Inc., 1968. 24. Dintentass, Leopold, "Blood rheology in ischemic diseases," Proceedings of the 8th International Conference on Medical and Biological Engineering, Session 1-1, July 1969. 25. Baker, D. W. and Strandness, D. E., "Instrumen- tation for early detection of arterial occlusive'disease," Proceedings of the 8th International Conference on Medical and Biological Engineering, Session 1-8, July 1969. 26. Cohen, M. V., Cohn, P. F., Herman, M. V., and Gorlin, R. G., "Diagnosis and prognosis of main left coron- ary artery obstruction," Supplement I to Circulation, Vols. XLV and XLVI, p. I- -57- I- 65, May 1972. 27. Dunn, F. and Fry, F. J., "Ultrasonic threshold dosage for the mammalian central nervOus system," IEE Trans- action on Biomedical Engineering, Vol. BME-18, No. 4, p. 253-256, July 1967. 28. Ulrich, W. D., "Ultrasound dosage for experi- mental use on human beings," Naval Medical Research Inst., 29. Hamilton, Andrew, "Seeing your ailments with sound," Today's Health, Vol. 46, p. 20-24, March 1968. 30. Franklin, D. L., "Techniques for measurement of blood flow through intact vessels," Med. Electron. Biol. Engng., Vol. 3, p. 27-37, Pregman Press, 1965. Printed in Great Britain. - ‘ 31. Kalmus, H. P., "Electronic flowmeter system," Rev. Scient. Instrum., 25, p. 201- 206, 1954. - 32. Haugen, M. G., Farral, W. R., Herrick, J. F., and Baldes, G. J., "An ultrasonic flowmeter," Proc. Nat. Electron. Conf., II, p. 465-475, 1955. . 33. Baldes, G. J., Farrall, W. R., Haugen, M. C., and Herrie, J. F., "A forum on an ultrasonic method for measuring the velocity of blood," p. 165-176, Ultrasound in Biology and Medicine (ed. E. Kelly) A.I.B.S., Washington, 1957. 90 ~34. Franklin, D. L., Baker, D. W., Ellis, E..M.,” and Rushmer, R. F., ”A pulsed ultrasonic flowmeter," I.R.E. Trans, Med. Electron. Me-6, p. 204-2-6, 1959. 35. Franklin, D. L., Baker, D. W., and Rushmer, R. F., "Pulsed ultrasonic transit time flowmeter," Ins. Rad. Engrs. Trans. Bio-Med. Electron, BME-9, p. 44-49. 36. Farrall, W. R., "Design considerations for ultrasonic flowmeters," I.R.E. Trans. Med. Electron. ME—6, p. 198-201, 1959. ‘37. Zarnstroff, W. D., Castillo, C. A., Crumpton,l C. W., "A phase shift ultrasonic flowmeter," Inst. Rad. Engrs. Trans. Bio—Med. Electron. BME-9, p. 199, 1962. 38. Nobel, F. W., Goldsmith, T. C., Waldsburger, C. J., and Cook, T. E., "A new system for measurement of blood flow by ultrasonics," Digest of 15th Conference on Engineering in Medicine and Biology, Chicago, 1962. ”39. wetterer, E., "A critical appraisal of methods of blood flow determination in amimals and man,“ I.R.E. Trans. Med. Electron. Me-9, p. 165-173, 1962. 40. Arts, M. G. J., and Roevors, J. M., "0n the instantaneous measurement of blood flow by ultrasonic means," Med. and Biol. Engng., Vol. 10, p. 232-34, 1972. 41. Green, P. 3., "Spectral broadening of acoustic reverberation in doppler shift fluid flowmeters," J. Acoust. SOC. Am.’ -36' P0 1383-1390, 19640 42. Satomura, 8., "Study of the flow pattern in peripheral arteries by ultrasonics," J. Acoust. Soc., Japan, Vol. 15, 1959, p. 151-158. '43. Franklin, D. L., Schlegel, W. A., and Rushmer, R. F., "Blood flow measured by dOppler frequency shift of backscattered ultrasound," Science, Vol. 132, 1961, p..564- 565. 44. Baker, D. W., et al., "A sonic transcultaneous blood flowmeter," Proc. 17th Am. Conf. on Engng. Med. Biol., 1964, p. 76. 45. McLeod, F. D., "A doppler ultrasonic physiologi- cal flowmeter," Proc. 17th Am. Conf. on Engng. Med. Biol., -l964, p. 81. 91 46. McLeod, F. D., "A directional doppler flowmeter,” Digest of the 7th Conf. Med. Biol. Engng., Stockholm, Sweden, 1967. 47. Yoshitoshi, Y., Machii, K., Sekiguchi, H., Mishima, Y., Ohta, 8., Hanaoka, Y., Kohashi, Y., Shimuzu, S., and Kuno, H., "Doppler measurement of mitral valve and ventricle wall velocities," Ultrasonics, Vol. 4, 1966, p. 27-28. 48. Watson, B. W. (Editor), "IEE medical electronics monographs," Peter Peregrinus,Ltd., 1971. 49. Barnes, R. W. and Thurstone, F. L., ”An ultra- sound moving target indicator system for diagnostic use," IEEE Trans. on Bio-Med. Eng., VOl. 18, January 1971, p. 428. 50. Flaherty, J. J. and Strauts, E. J., "Ultrasonic pulse doppler instrumentation," 8th International Conference on Medical and Biological Engineering, Chicago, 1969. 51. Peronneau, P. A. and Leger, F., "Doppler ultra- sonic pulsed blood flowmeter," 8th International Conference on Medical and Biological Engineering, Chicago, 1969. 52. Hokanson, D. E., Mozersky, D. J., Summer, D. S., and Strandness, D. E., "Ultrasonic arteriography and phase-, locked echo tracking system for measuring arterial diameter changes," IEEE Ultrasonic Symposium, Miami Beach, 1971. 53. McLeod, F. D. and Anliver, M., "Multiple gate pulse doppler flowmeter," IEEE Ultrasonic Symposium, Miami Beach, 1971. 54. Peronneau, P., Pellet, M., Burgon, J., Xhaard, J., Hinglasis, N., Latil, F., and Couillard, F., "An ultra- sonic multigated pulsed Doppler velocimeter for real time measurement of blood velocity profiles," 25th Annual Con- ference on Engineering in Medicine and Biology, pp. 180, 1972. 55. Skolnik, M. E., "Introduction to radar systems," McGraw-Hill, 1962. . 56. Wagg, R. C., and Gramiak, R., "Noninvasive- heart blood flow rate measurement," Presented at 16th Annual Meeting of the American Institute of Ultrasound in Medicine, Denver, Colorado, 19-22 October 1971. 92 57. Wagg, R. C., Rhoads, W. L., and Gramiak, R., "Instrumentation for noninvasive cardiac chamber flow rate measurement," IEEE Ultrasonic Symposium Boston October 4-7, 1972. 58. McGillem, C. D., Cooper, G. R., Waltman, W. 8., "Use of wideband stochastic signals for measuring range and velocity," EASCON Conference Record, pp. 305- 311, 1969. 59. Cooper, G. R. and R. L. Gassner, "Analysis of a wide band random signal radar system,” Purdue University, 60. Cooper, G. R. and C. D. McGillem, "Random sig- nal radar,” Purdue University, TR-EE-67-11, June, 1967. 61. Cooper, G. R. and R. J. Purdy, “Detection, Resolution, and accuracy in the random signal radard, ' Purdue University, TR-EE- 68- 16, August, 1968. 62. C. D. McGillem and W. B. Waltman, ”An experi- mental random signal radar system," Purdue University, TR- EE-69-46, December, 1969. 63. Rihaczek, Principle of High-Resolution Radar, McGraw Hill, 1969. 64. Newhouse, V., Cooper, G., Feigenbaum, H., Jethwa, C. P., and Kaveh, M., "Ultrasonic blood velocity measurement using random signal correlation techniques.” Invited for a presentatiOn at the 10th International Conference on Medical and Biological Engineering, Dresden, GDR, August 13-17, 1973. 65. Egli, P. H., "Ultrasonic beam transducer," United States Patent Number 786,258, December 1968. 66. Wilkox, M. H., and Egli, P. H., "Ultrasonic- Amplitude-Dopper Detector," United States Patent Number 787, 397, December 1968. . - 67. Jethwa, C. P., Saggio, Frank III, “Resolution of a pulsed random signal ultrasonic doppler system." Accepted for a presentation at the 18th annaul meeting of the American Institute of Ultrasound in Medicine, Ann Arbor, Michigan, October 14-18, 1973. 68. Chadwick, Russell 3., "Use of random signals to study distributed radar targets," Ph.D. thesis, School of Electrical Engineering, Purdue University, August 1970. 93 69. L. A. Geddes,."Measurement of blood pressure," Year Book Medical Publishing, Inc. 1970. 70. McCutchen, E. P. and Rushmer, R. F., "Korotkoff Sounds,” Circulation Res. 20:149-161, 1967. 71. Tursky, B., Shapiro, D., Schwartz, G. E., "Auto- mated constant cuff-ressure system to measure average systolic and diastolic blood pressure in man," IEEE Transaction on Bio- medical Engineering, Volume RME-19, Number 4, pp. 271-276. 72. Sigel, 8., Gibson, R., Amatneek, K., Felix, R., Edelstein, A., Popky, G., "A DOppler ultrasound method for distinguishing laminar from turbulent flow," Journal of Surgical Research, Vol. 10, No. 5, pp. 221-224; May 1970. 73. Satomura, S., "Ultrasonic Doppler method for the inspection of cardiac functions." J. Acoust. Soc. Am., 74. Lube, V. M., Safonov, Yu. D. and Yakiemenkov, L. I., ”Ultrasonic detection of motions of cardiac valves and muscle." Soviet Phys. Acoust. 13, pp. 59-65, 1967. -. I‘llll'l' il‘.lilil| (illl‘ fill-I‘ll “I? flilfififlgfigfimflfimfltu 11111717117171 ES 838731