3. THE NON-INVASIVE ASSESSMENT OF UTERINE ACTIVITY J. Nagel and M. Schaldach Zentralinstitut fur Biomedizinische Technik der Universitiit Erlangen-Nurnberg, Erlangen, FRG 1.0 INTRODUCTION The monitoring of uterine activity is of importance in prenatal medicine for the surveillance of the course of the pregnancy, and for assessing the condition of the fetus. The relevant information about uterine activity can be derived from mechanical effects, such as deformation, tension in the abdominal wall, changes in intrauterine pressure, and from electrical phenomena produced by the uterus, i.e. the muscle potentials which trigger the mechanical activity. Since the uterus is not readily accessible for the direct recording of the mechanical aspects of its activity, the results of such measurements, as performed clinically, provide only little information about the state of the uterus. Normally, the activity is evaluated in terms of the intrauterine pressure, usually in the form of the so-called tocogram. Two methods, one invasive and the other non-invasive, may be employed. For internal (invasive) tocography, the pressure is determined by introducing a catheter-tip pressure-transducer, or a catheter connected to an external transducer, into the uterine cavity. With this method, it is possible to determine the absolute intrauterine pressure. Often, however, internal tocography is not employed because of the practical difficulties involved. Moreover, the insertion of a catheter may stimulate parturition, so that this method may be contra-indicated for possible premature deliveries, NON-INVASIVE MEASUREMENTS: 2 Cop}'Tlght©J983 by Academic Press Inc. (London) Ltd. ISBN 0 125934025 AU rights of reproduction in any form reserved 104 J. NAGEL AND M. SCHALDACH 1.0] where monitoring of the uterine contraction might be of particular clinical value. External (non-invasive) tocography, in which intrauterine pressure changes are inferred from measurements on the abdominal wall, avoids these disadvantages of the invasive measuring techniques. However, it has the drawback of being only a relative measurement, and is also subject to numerous interfering influences which may severely affect the results it provides. Although invasive and non-invasive tocography have, until fairly recently, been the only methods employed for the assessment of uterine activity, it has been shown that they do not allow detailed information about the mechanical activity of the myometrium to be determined. Simultaneous recordings of intrauterine and intramural pressure reveal local increases in wall tension that are not always mirrored by the intra- uterine pressure. Since the fundamental problems of tocography cannot be eliminated by technical improvements to the measuring equipment, it would seem justifiable to consider assessing uterine activity by processing the uterine electrical action potentials, which are directly related to the activity of the myometrium. As clinical trials have shown, this technique yields much more information, in particular with respect to the excitation and propa- gation of the contractions. 2.0 TOCOGRAPHY The procedure of recording uterine activity can be tr.1ced back to the period around 1870. At that time, Kehrer (1867) and Schatz (1872), published the first ever intrauterine pressure curves which, in their precision and quality, remain exemplary even today. They measured the intrauterine pressure by means of a liquid-filled balloon catheter, which they had introduced between the uterine wall and the membranes. The disadvantages of invasive pressure recording, in particular the associated danger of infection, precluded its use in the clinical setting for a long time. Since internal (invasive) tocography is still relatively rarely employed, and since this article is limited to non-invasive techniques, this method will not be dealt with further here. External (non-invasive) tocography, which still remains the most commonly employed method for the monitoring of uterine activity, can also be traced back to the last century. In the obstetrical literature, reference can be found to an initial attempt to carry out external tocography using a large, air-filled metal box made by Schaffer in the year 1896. Subsequently described external devices for the measurement of uterine contractions, 2.0] 50 40 30 Zero point: 20 external tocography ---'0 3. ASSESSMENT OF UTERINE ACTIVITY Controction omplitude (intensity) lOS Basal pressure toc~t~~n~~- ol-----.r----.----r--....----r---r--'-- 50 100 200 300 t (s) FIG. Schematic representation of the course of uterine contraction. With internal tocography, pressure and intensity are recorded in absolute values, external tocography providing only relative values, without the possibility of determining the basal pressure. In the monitoring of uterine activity, apart from pressure, the form of the contractions also plays an important part. such as the tocodynamometer (Crodel, 1927) and the external hysterograph (Riibsamen, 1920) had considerable shortcomings, but showed that there was very early interest in obtaining a continuous, external measurement of uterine contractions. Further progress was made by the development of a hysterotonograph (Frey, 1933) and a tocograph (Lowi, 1933), but they required that the entire recording unit be affixed to the abdomen of the bell Fetal membranes 2 3 FIG. 2 Common methods of tocometry. (1) External tocometer with "tocometer pin", which is affixed to the abdomen by means of a strap. (2) Intrauterine, extra- amniotic method using a fluid-filled balloon catheter. (3) Trans-cervical intra- amniotic pressure measurement using an open-end catheter. 106 J. NAGEL AND M. SCHALDACH [2.0 mother. An electromechanical uterine contraction recording device that is relatively insensitive to interference has been known since the work of Rech (1934). The uterine contraction transducer is attached to the abdo- men by means of a rubber strap, and the recording device is set up at a distance from the patient. The transducers in common use today still employ the principle described by Rech. Contractions of the uterus cause a sensing pin con- tained within the transducer to be deflected mechanically from its resting position, and the deflection is detected by a strain gauge. A number of strain gauges are connected together to form a measuring bridge (full or half-bridge). The change in resistance of the strain gauge accompanying a mechanical deformation is converted into an electrical signal, which represents a measure of the deformation and thus of the strength of the uterine contraction. Despite very thorough investigations, agreement has still not been reached as to what, in fact, this transducer is measuring. The "hardness" of the uterine wall is a complex parameter which depends upon the wall tension, radius, wall thickness, internal pressure, transverse elasticity and the hardness of the uterine musculature, and is just as involved in the measurement, as is the deformation of the uterus and its "erection" during contractions. A further component of the measured signal results from the elastic nature of the transducer attachment. It can thus readily be appreciated that external tocography does not permit any statement about the absolute contraction amplitude nor the absolute level of the basal pressure or tone. All that we can obtain is an impression of the relative intensity of uterine contr~ctions and changes in the basal pressure. Even this restricted performance may only be obtained when it is ensured that the transducer remains at its original site throughout the recording period and that the measuring conditions are not changed by any alteration of the patient's position. In very many cases, these requirements cannot be met; indeed, the maintenance of a given position by the patient might even be dangerous for her and the fetus. In general, external tocography can reliably reproduce contraction rate and approximate form. Movements of the fetus can be recognized as small arrhythmic peaks, while respiratory movements appear as super- imposed rhythmic waves. Care must be exercised, however, in the identi- fication of the movements of the fetus since similar peaks in the contraction curve can also be produced by. brief contractions of the abdominal wall musculature. Although tocography has become a routine method for the monitoring of uterine contractions, and, today, almost no pregnancy remains without tocographic monitoring, it is not capable of providing information on 3.0] 3. ASSESSMENT OF UTERINE ACTIVITY 107 the detailed mechanical activity of the myometrium. Cibils and Hendricks (1969) made simultaneous recordings of intrauterine and intramural pressures by means of open-tip catheters inserted into the uterine cavity and the myometrium, respectively. They observed local increases in wall tension, which were not always accompanied by measurable increases in intrauterine pressure. These measurements were made in the post-partum uterus, but this does not invalidate the conclusion that a recording of intrauterine pressure is not a faithful reproduction of detailed activity of the uterine musculature. 3.0 THE ASSESSMENT OF UTERINE ACTIVITY ON THE BASIS OF ITS MYO-ELECTRIC SIGNALS The myo-electric signal is the electrical manifestation of any contracting muscle. It seems reasonable to assess the possibility of obtaining accurate information about the mechanical activity of a muscle by analysing its myo-electric signals (electromyography, EMG). Many attempts have been made to record and analyse the electrical activity of the pregnant uterus. The lack of suitable measuring and signal-processing methods, and the resulting poor, or even false, results, meant that the possibilities of electromyographic monitoring of uterine activity long remained unrecog- nized. Only recently has a procedure been described (Nagel and Schaldach, 1980a) which resolves the earlier problems and permits the reliable deter- mination of uterine activity on the basis of its myogram. 3.1 The Electrical Activity of the Uterus Numerous attempts have been made to record the electrical activity of the uterus. Larks (1960) and Wolfs and van Leeuwen (1979) published a detailed review of the historical development in this area of research. In the majority of cases, the myo-electrical signals were picked up via skin electrodes affixed to the abdomen; in a number of cases, the EMG was recorded invasively using needle or micro-electrodes. A wide variety of different measuring methods and equipment were employed. Thus, it is not surprising that the numerous investigations produced widely varying results. The signals measured were simply attributed to the uterine activity without any attempt to check their actual origin. There are, of course, numerous possible sources of artefacts, such as movement of the patient, including respiratory movements, the electrical activity of the muscles of the abdominal wall, the smooth muscles of the intestine and the bladder, the electrocardiogram of the mother, skin potentials and movement 108 J. NAGEL AND M. SCHALDACH [3.] artefacts caused by the contracting uterus, all of which tend to degrade the signal-to-noise ratio. Bode (1931), Clason (]934) and Mestwerdt (]944) recorded slow, biphasic waves, with faster fluctuations superimposed upon them. They suggested that part of the activity they had recorded was due to the heart, or the mechanical or electrical activity of the respiratory and abdominal wall muscles. Many other investigators, such as Dill and Maiden (] 946), Steer and Hertsch (]950) and Halliday and Heins (]950) also recorded very low-frequency electrical signals (0·] -2 Hz), their measurements varying greatly with respect to the shape and occurrence of the signals. Muller and Liechty (]954) discovered more electrical activity between contractions than during contractions. Steer (1954) described two different types of electrical activity during uterine contraction: slow waves having a periodicity of several seconds, upon which faster waves (0·3-2 Hz) were superimposed. Sureau is one of the leading investigators in this field (Sureau, 1955, ]956, ]964; Sureau et al., 1965). During contraction, he recorded sinusoidal waves having a frequency of 0·3-1 Hz. Larks (1956) obtained results similar to those of Steer. The amplitude of the electrical signals measured by different investigators varies considerably, covering a range of between 50 (lVand ]50 mV. Very extensive investigations were carried out by Wolfs and Leeuwen (1979), and although they did not succeed in obtaining all of the information contained within the uterine EMG, their measurements did reveal a marked correlation between the EMG and intrauterine pressure. A very relevant question is why so many working groups recording the uterine myopotentials have obtained such a wide variety of different results. The reason for this would appear to be that the investigations have been based on incorrect or incomplete theoretical models of the origin, propagation and measurement of the electrical signals, and on the mode of electromechanical coupling. The result of this was the use of unsuitable measuring equipment for the recording of the signals. Thus, for example, in many cases, DC-coupled or very low-frequency measuring amplifiers were employed because adequate attention had not been paid to the problem of temporarily changing electrode potentials. The frequencies of these changes are in the same frequency range as the signals under investigation and can almost completely mask the useful signal. Thus, the question as to the frequency spectrum of the signals was completely ignored, and as a result, most measurements recorded only the fluctuating resting potentials, but not the action potentials typical for the activity of the musculature. Furthermore the selection of the most suitable combination of electrodC$ and recording positions and of the origin and composition of the signals, was not subjected to a systematic examination. The resolution of these 3.1] 3. ASSESSMENT OF UTERINE ACTIVITY 109 problems is, however, an essential pre-condition for the careful analysis of the electromyographic signals picked up from the uterus. An under- standing of the physiology of labour presupposes a knowledge of a number of basic biological principles. Here, therefore, the fundamental processes in the formation of bio-electrical potentials, and electromechanical coupling, are briefly described. Every animal cell is bounded by a highly differentiated membrane (cell membrane) which regulates the exchange of substances between the intracellular and extracellular spaces. This membrane has the capability of being selectively permeable and effecting active transport. The intra- cellular and extracellular spaces differ in their ionic concentrations. Within the cell, the concentration of potassium ions is 40 to 50 times as high as that on the outside. Sodium ions, on the other hand, have an extracellular concentration 3 to 10 times that of the intracellular concen- tration. Owing to this difference in ion concentrations, an electrical potential difference develops between the inside and outside of the cell. The resting potential of the cell membrane is between -60 and -90 mY. In the resting state of the cell (polarized), the intracellular potential is negative with respect to the extracellular space. Nerve and muscle cells are characterized by the fact that stimulated or autonomous activation can have an effect on the cell membrane. As a mV Depolorlzotlon Aepolorlzohon t- -t- --l Overshoot + 20 - - - - iin(t), which, of course, is dependent upon time. The force J(t) can then be expressed thus N J(t) = C1 • L gn(t) . 6>n(t). (12) n-I If it be assumed that the individual muscle fibres are excited synchronously, and, further, that they each develop an identical contraction force, then we obtain, for the force J( t) J(t) = Ct . h(t) . 6>(t), (13) in which h(t) is the numbe~ of activated muscle fibres. The question must now be examined as to whether the intensity of the EMG can be expressed in a similar manner, with the aid of the stimulation frequency. For this purpose, Eqn (6) is put into a more easily interpretable form. We shall first consider the case in which the measuring point and the co-ordinate origin coincide-in the centre of the spherical theoretical uterus. Here, r = O. Equation (6) reduces to: N _ U (t) = 2J 4!cr f - div ~n (r', t) dV' * ait) }. (14) II-I V' The point of departure for further simplification of (14) is the vector identity: v· (a/r) = (V' (;)/r + (;. V (l/T). (15) The integration of all three terms in volume V, which contains all sources, provides us, on applying the divergence theorem to the first term, with the 3.4] 3. ASSESSMENT OF UTERINE ACTIVITY 117 equation (16) s v v Since G = 0 on the entire limiting surface area S, it follows from (l6) that (17) v v So that (14) can be re-formulated as follows: N U(t) = L{4!aJe. (;-, t)·V(:,)dV' *a.(t)} (18) """,,1 V' or ~ - Vet) = ""{_I_J - 6.(;',t)· ~dV'*a.(t)}. (19) L 41ta r'3 n=1 V The flow of ions, G, is a source function, which is interpreted as a dipole moment per unit of volume. A contribution to the potential at the measur- ing point is made only by the !!ldial (Gn,), but not by the tangential, com- ponents of the dipole vectors G •. Accordingly, the following equation N _ Vet) = ""{_I_J- G."V, t) dV' * a.(t)} L 41ta r 2 (20) applies. With the assumed spherical symmetry and the relatively small thickness of the myometrium, for an estimation of the potential the contribution of the radius vector can be considered constant for all fibres, so that the factor l/r'2 can be removed from the integral and the sum. Under these pre- conditions, the integral can be expressed as a functionYn(t), now depending only on time. Equation (20) becomes: N Vet) = I , "" Yn (t) • a.(t). 41tar 2 L (21) With synchronous stimulation of all muscle fibres, and the same shape of the curve of all Yi(t), because of the refractory period of the cells, no 118 J. NAGEL AND M. SCHALDACH [3.4 interference phenomena can occur. In this case, the potential can be expressed by U (t) = 1 k· (t)· (y(t) * a,,(t». 41t(J r'2 (22) where k(t) is the number of muscle fibres stimulated. If the potential function described in (7) is rectified and filtered, applying the same arguments as for the strength of the mechanical contraction, the spike train a(t) can be replaced by the stimulation frequency w (t) and we obtain, with constant C3 for the intensity of the EMG, an expression having the form: I (t) = C3 • k(t) . wet). 4r:cr r'2 (23) A comparison of (23) and (13) shows that for constant r', the force of contraction and the intensity of the EMG (IEMG) are proportional to each other: let) = c4 • l(t). (24) In addition, using (11) pet) = Po + cs· l(t). (25) applies. Accordingly, in the special case under consideration, both the intensity of contraction and the relative intrauterine pressure can be determined from a measurement of the myo-electric potential. Of course, the question may now be asked as to what practical signifi- cance this result has. Although, in the derivation of (24) and (25), so many approximations were made that one might not expect the result to be quantitatively correct, experimental investigations show that at least qualitatively it does in fact conform to the physiological situation. The reason for including this derivation here, however, is to make it clear that, at least in principle, it is possible to find a fixed relationship between the EMG, and the intensity of contraction of the uterus. This would seem all the more important since, to date, reports in the literature have all denied this possibility. This is possibly the result of the fact that, formerly, it has always been the EMG itself, but not its intensity, that has been evaluated. We must now investigate the nature of the relationship between J(t) and l(t) without the restricting conditions assumed in the derivation of (24) and (25), in particular in the case of an external recording of the myopotentials. In this connection, we must first examine the question as to whether the dependence of the intensity of the EMG on the force of 3.4] 3. ASSESSMENT OF UTERINE ACTIVITY 119 contraction of the muscle changes when the assumption of synchronous stimulation of the muscle fibres-which is certainly not really the case-is dropped. For the mathematical determination of the relationship then applicable, the statistics of muscle excitation, and the geometry of the individual muscle fibres, must be known. On account of the complexity, perhaps even the impossibility, of this computation, no attempt was made to adopt this approach. Instead, experiments were performed to discover whether the linear relationship betweenf(t) and l(t) is preserved. According to the literature (Person and Libkind, 1967; De Luca and Forrest, 1973), this is not the case for all muscles; sometimes, there is a square law relationship (/(t) oc /2(t) ). With respect to the myometrium, however, we have been able to confirm the linear relationship already found by Milner-Brown and Stein (1975) for a number of other muscles. Accordingly, therefore, the assumption of synchronicity of fibre stimula- tion made in the derivation of the relations (24) and (25), does not result in a qualitative falsification of the result. A further simplification, whose influence on the results has to be investigated, is the assumption of a spherical uterus and potential measure- ment in the centre of the sphere. Only under the above-mentioned conditions do the contributions of the individual dipole vectors in the overall potential, have identical weight. Both in the case of potential measurement outside of the centre of the sphere, and also in a change in the geometry of the uterus, a non-uniform weighting of the individual sources results. In accordance with Eqn (6), the influence of such sources that are closer to the measuring point becomes more marked, while those potentials originating in more distant muscle fibres become more attenua- ted. The result of this is that on moving the measuring point to a given part of the myometrium, mainly the activity of the muscle fibres in the immediate neighbourhood is recorded. Thus, by appropriately siting the electrodes, the local activity of individual regions of the uterus, or the spread of the contractions can be picked up. In this manner, motility disturbances, such as incoordination for example, can also be recognized. For the global, non-invasive determination of the uterine activity, the measurement of the myopotentials at a single point is not adequate. For there is no point outside of the uterus that is equidistant from all the muscle fibres. Owing to the large spatial extension of the uterus, this condition is not even approximately fulfilled. Nevertheless, the uterine activity can be determined globally, if the potential is picked up simul- taneously at a number of points. By appropriately siting the measuring electrodes and summating the individual potentials a measuring signal is obtained which can be used in Eqns (24) and (25) to give adequately accurate results. The degree of this approximation depends upon the 120 J. NAGEL AND M. SCHALDACH [3.5 number of measuring points and on their sites. Clinical investigations have shown that, in the majority of cases, the measurement of potentials at two points on the maternal abdomen is sufficient to ensure a result that is representative for the whole uterus. 3.5 Interference Potentials The external recording of the uterine EMG is made difficult by super- imposed strong interference signals arising in the maternal ECG, the fetal ECG and the EMG of the abdominal wall musculature. The amplitudes of the interference signals are usually greater than those of the useful signal. Further possible interference components, such as electrode offset potentials, electromagnetic interference and noise potentials, are not considered here, since they can be avoided or suppressed by designing suitable measuring systems. In passing it might be mentioned that the bio-electrical interference signals are not included in the measurements indicated in the literature. One reason for this is, that in most measure- ments reported bipolar electrodes were used. With these the potential difference between two electrodes located close together is determined, so that the contributions to the signal of more distant sources, e.g. the heart, are strongly attenuated. The second reason is that the frequency response pf the amplifiers employed only allowed signals below 2 Hz to be measured. In this low-frequency range, however, the interference signals mentioned have only a very small power density, so that their contribution to the measured signal is small. An analysis of the frequency spectrum of the uterine EMG, however, shows that it extends to about 250 Hz and has its greatest power density above 2 Hz. Accordingly, the recording of the EMG should be carried out in this frequency range. The low-frequency range below 2 Hz is not suitable for routine measurement since it is here that movement artefacts are strongly seen. The use of closely spaced bipolar measuring electrodes is reasonable only for the pick-up. of local muscle activity. For the analysis of the intensity of the EMG, the interfering components must be suppressed prior to rectification and low-pass filtering. The maternal ECG can be subtracted from the original signal using a pro- cedure described by Nagel and Schaldach (1980b). The R-peaks of the maternal ECG (MECG) are easily detectable by means of threshold detectors on account of their prominence in the abdominal signal. Through the exponential averaging of succeeding segments of the signal, all con- taining the maternal QRS complex in the same phase position, a reference signal corresponding to one interval of the MECG is obtained. The other signal components are suppressed in the reference since they are statistically 3.5] 3. ASSESSMENT OF UTERINE ACTIVITY 121 independent of the MECG. Subtraction of the reference from the abdominal mixed signal after a special scaling operation, results in the complete elimination of the MECG. The fetal ECG can be eliminated in the same manner although, as practical experience shows, this is not necessary because of its small amplitudes; its influence on the labour (uterine activity) curve is negligible. A simplification of the procedure is achieved by limiting the potential measurement to the frequency range from c. 150 to 250 Hz. Since, here, the amplitude of the uterine EMG is considerably greater than that of the fetal and maternal ECG, signal separation is not necessary. When measurements were carried out in this restricted frequency range, no changes were observed in the labour curve. The only interfering component that cannot be eliminated from the mixed signal, but can merely be reduced by appropriately positioning the electrodes, is the EMG of the abdominal wall musculature. This fact, however, is not necessarily a disadvantage. Its contribution to the labour curve is so characteristic that it can clearly be distinguished from the intensity of the uterine EMG. Furthermore, it can also show the behaviour of the mother under the stresses of labour, e.g., during expulsive con- tractions. Over and beyond this, it can be observed that the contractions of the abdominal wall muscles also lead to an increase in intrauterine pressure. Thus, the additional recording of the activity of the abdominal wall musculature is desirable rather than undesirable for the practical application of the measuring procedure described. In any case, the contractions of the abdominal wall muscles also strongly affect the external mechanical pressure recording. 3.6 Movement Artefacts According to Eqn (6), the uterine myo-electric potential is dependent upon the measuring point. Thus it is to be expected that movements-either changes in the position of the patient, or changes in the geometrical state occurring during uterine contractions-will influence the potential of a measuring electrode affixed to the maternal abdomen. Depending upon the movement and the position of the electrode, or on the transmission path of the bio-electric signals, their amplitude decreases or increases, so that this effect also modulates the EMG, and can thus falsify the labour curve. The error can be eliminated by having the signal amplifier com- pensate for the fluctuations in amplitude. It is, of course, not desirable to adjust the amplitudes of the EMG to a constant level, since important information about its intensity would then be lost, for the amplitude changes caused by contraction would also be eliminated. 122 J. NAGEL AND M. SCHALDACH [3.6 :-100 100 --- I---f-- -- _.- I---~~5 11\ 7~_ -I-- II \ Ir".. \. - - I---5 \. J \ .,..... ! I f I H J \ ./ V ] "\ r \. ./ / \. 2f \. ..... V v "" "').. "- ./ ...., tlllOmOilU( IEMG with amplitude adjustment t_ r-Iyv 100 I min .1--- - 1----1- -- ---I--- t--7 7r_ ~~ ---.-I---5p r-25 ~ -'~ - .....:;; h ./ I-..... I---' ,... A tBlOY'RONIK IEMG without amplitude 'bdJustment eBJO'i'N)MICC FIG. 4 Influence of amplitude adjustment of the El\IG on the uterine contraction curve. A surprising but simple solution is the use of the maternal ECG as a calibration signal for amplitude control. Experimental experience has shown that the amplitude fluctuations of the MECG correspond quite accurately to those of the EMG. From this it may be deduced that possible measuring errors can be avoided by employing automatic gain control that ensures the constant amplitude of the MECG within the original signal. In Fig. 4, the effects of adjusting the amplitude of the EMG to the contraction curve is represented by simultaneously recording the contrac- tion curve with and without amplitude adjustment, using the same pick-up electrode. 4.0 COMPARISONS OF RECORDINGS OF THE UTERINE ELECTRICAL AND MECHANICAL ACTIVITY Below, a number of examples of the recording of uterine activity are described, and these are intended to show the degree of conformity and also the differences between tocography and the recording of uterine activity via the EM G. There is overall good agreement between the theory and actual measure- ment. Figure 6 shows a comparison of contraction curves measured mechanically and myographically. For the detection of the EMG, an electrode was placed at the isthmus and another at the fundus of the uterus. The potential reference point was obtained from a third electrode applied to the thigh. In contrast to the externally measured pressure (upper curve), Is ol at io n A m pl ifi er E le ct ro de s a m pl ifi er a u to m at ic o a ln 1- 25 0 Hz c o n tr ol (a ) R -w av e de te ct or I- Is ol at io n A m pl ifi er E le ct ro de s o m pl ifi er a u to m at ic 0 0' " f - - ~ 1- 25 0 Hz c o n tr ol (b) R -w o v e de te ct or t- - Fi lte r 15 0- 25 0H z F ilt er 15 -2 50 H z ! R ec og ni tio n M EC G FE CG M HR FH R R ec tif ie r R ec tif ie r f E lim in at io n M EC G Lo w -p as s fil te r II t ) 0- 1 Hz Lo w -p as s fil te r I (I ) 0- 1 Hz FI G . 5 B lo ck d ia gr am s o f t he e le ct ro ni c c ir cu it fo r re c o rd in g el ec tri ca l u te ri ne a c tiv ity . ( a) T he s im pl er a rr a n ge m en t, in w hi ch o n ly th e hi gh fr eq ue nc y c o m po ne nt s o f t he E M G a re p ro ce ss ed . ( b) T hi s c ir cu it pe rm its th e re c o rd in g o f t he la bo ur cu rv e fr om th e c o m pl et e E M G a n d al so th e m a te rn al a n d fe ta l h ea rt ra te s fr om th e re sp ec tiv e E C G s. B ef or e pr od uc in g th e e n v e lo pe s o f t he E M G . t he m a te rn al E C G is e lim in at ed . " ' " ' , " " ,. M ''''' '",' 1 1'" , " ' , '/" f ', I~!,, ''''' ''!'I ' I" 'r w'i~ *,' IV "' I'!, 1 :" :," _: It ",' r J . , ,,/' :~ ' i't, :.l il: ' " ," ,I'~ j, I. : /11 ,\ :" ,III I ' I' ~,l ',. <,. , ",' ,j" ( • . , ~Ld " . . ~ .. 1 J "., .•. ~.,\ •. ~ ... ,,~ j.", ,,~, . . , ," . " ,~, ,\ . " " . . " . , ', ," ,, ,, ' • . , ' ~~, ~, ' j ( a ) . . . I m in Ar t No 12 0 0 1 0 A rt . No .1 2 - - - + - t r"~ I \ I . \ - - _ - - '° to : ~ _ - - - - ~~ I ~O=- - f\ - - - \-- -~ - - - - - - 1-- --~~ ~-~- -\- - - - - - - - 1 - - - 71 .. I.'V 1-\ Vi - - I--- .Y - f- r- -- . - - \ - 1-1 \- . 1 ~ - f- ~I -\ -- ~ . A : 'f' H ~ - - \ \ ~~ Ir -I \ f-- j bIi' -\ I \ ',0 , - \.k r 1" '- 25 - \ . ~ -~. .," " It. \. 15 - ~ - - - \ II \ . . . . 1 r I" 'U . , - 'V iJ/ ~ 1 '- ' ~ . ~ r - - - - ,. A I' " " ~. J _ ), ° ~ ~ 81 11 Om Oi 11 K • BI O'Y 'N)I 1!') ICC e lll O m O ili K FI G . 6 U te ri ne c o n tr ac tio n (la bo ur ) c u rv e , re c o rd ed s im ul ta ne ou sl y w ith a n e x te rn al p re ss ur e pi ck -u p (a) a n d fr om t he e le ct ro - m yo gr am (b ). .. "0 ~------- a. E