8/13/2019 Features of Voltage Pulse Plethysmography

1/6

FEATURES OF VOLTAGE PULSE PLETHYSMOGRAPHY

Martina Melinak, B.Sc., Polytechnic of Karlovac, 47000 Karlovac, I. Metrovia 10, Croatia, martina.melinscak@vuka.hrAnte anti, Prof. D.Sc., Faculty of Electrical Eng. & Computing, 10000 Zagreb, Unska 3, Croatia, ante.santic@fer.hr

ABSTRACT

Measurement of tissue volume changes by voltage pulseplethysmography is being studied. In voltage pulseplethysmography short voltage pulses activate the bio-tissue and the characteristics of the transient process aremeasured. There are two possible implementations: withgrounded bio-tissue and with grounded current samplingresistor. Measurement was carried out for both cases andthe obtained wave-shapes for blood pulsations in thearteries and for respiration are presented. From thecharacteristics of measured voltage the parameters ofequivalent circuit for electrode-skin interface with three

elements are calculated for dry and gel-wetted electrodes.KEY WORDS

voltage pulse plethysmography, current pulseplethysmography, electrode-tissue interface, tissueresistance variation

Introduction

Impedance plethysmography is mostly used for longertime clinically. Pulse plethysmography is the object ofnewer researches. antiet al [1], [2] mention advantagesof current pulse plethysmography compared to impedance

plethysmography. Main advantage of applying short

current pulses in current pulse plethysmography instead ofsinusoidal current in the impedance plethysmography isthat in the pulse plethysmography the amplitude forsensations threshold is much higher then in impedance

plethysmography implying that better signal-to-noise ratiocould be achieved.

The purpose of this paper is to test characteristics ofvoltage pulse plethysmography, and to compare it withcurrent pulse plethysmography which has already beenexamined [1], [2].

Measuring instrumentation

There are two possible implementations of voltage pulseplethysmograph, with grounded bio-tissue and withgrounded current sampling resistor, R0 (Fig. 1).Measurement for both cases was done in order toinvestigate advantages and disadvantages of both cases.

um(t)

u 0(t) u 0(t)

um(t)

Fig. 1. Implementation with one electrode connected to

bio-tissue grounded (left) and with current samplingresistor grounded (right)

Fig. 2 shows the circuit diagram of the voltage pulse

plethysmography measurement set-up. The voltage pulseplethysmography set-up consists of the variable powersupply (voltage source) U0, the transistor switch T1, theswitch control digital circuit (A.M., M.M.1 and M.M.2),reset transistor switch T2, S&H circuit and the a.c.coupled amplifier with incorporated notch filter.

R0

U 0

u SH (t)

u out (t)u m (t)

Z X

A.M.

M.M.1

M.M.2

M.M.3

Notch-filter

50 Hz

T1

T2

T

C

R

Fig. 2. The block diagram of voltage pulse plethysmography

8/13/2019 Features of Voltage Pulse Plethysmography

2/6

The voltage pulses are generated by astable multivibratorA.M., monostable multivibrator M.M.1, and PNPtransistor that works as a switch which connects voltagesource U0 with very little internal resistance to the bio-tissue and the current sampling resistor. Transistor T2 isused as a switch that discharges the capacitive component(CX, Fig. 3) of electrode-skin interface impedance.Monostable multivibrator M.M.3 is controlling ON and

OFF transistor T2.The frequency of pulse repetition is defined by astablemultivibrator A.M. switching at 10 kHz. As Fig. 2 showsmonostable multivibrator M.M.1 is triggered by the

positive edge of astable multivibrator pulse. The ON timeof M.M.1 controls the duration of the voltage pulse (theON time of transistor T1). The negative edge of themonostable multivibrator M.M.1 pulse triggersmonostable multivibrator M.M.2 which controls S&Hcircuit. Positive edge of monostable multivibrator M.M.1

pulse triggers monostable multivibrator M.M.3, whichcontrols transistor T2 and which discharges bio-capacitance CX.

The measurement chain consists of S&H, high pass filter,notch filter for 50 Hz and amplifiers. The S&H samplesand holds measured voltage um after sampling time T,which is determined by M.M.2 monostable multivibrator

pulse duration. The high pass CR filter removes DCcomponent of the voltage um(t). High pass filter cutofffrequency is 0.05 Hz. The 50 Hz noise from power lines isreduced by the notch filter. The total gain of the amplifierchain is A=390. Also it should be mentioned, that voltage

pulse generator is much simpler than current pulsegenerator.

Electrode-skin interface

The electrode-skin interface can be modeled by a five-element circuit that can be simplified to a three-elementcircuit due to the circuit symmetry (Fig. 3) [2]. Theimpedance of the electrode-skin interface is frequencydependent [3]-[5] and at high frequencies [2], [6] can bereplaced by a serial RC-circuit, while resistance valueRSXis variable due to the change of the bio-tissue volume(blood pulsation, respiration, etc.). The resistancevariations are rather small, typically 0.1-0.5 % [1], [6].

Fig. 3. Equivalent circuits for electrode-skin interface [2]

Wave-shapes corresponding to blood

pulsations and respiration

It must be pointed out that maximal sensitivity UX/UXisobtained when series resistance of the sampling resistor isequal to the bio-resistance. It is valid for both cases: with

bio-tissue grounded and sampling resistor grounded.

Simple calculation can give us when maximal voltagevalue between systolic pressure voltage US and diastolicvoltage UD pressure UX=US-UD can be obtained incomparison to UX=UD i.e. UX/UX. Systolic pressurevoltage is UX+UX and diastolic voltage is UX. Thedifference is:

SXSXSXDSX

RR

UR

RRR

URUUU

+

+==

0

00

0

00

)(. (1)

Looking for extreme value it can be found that it occurswhen bio resistance RSX is equal to current samplingresistanceR0.Fig. 4 shows the measured voltage um(t) while measuring

pulsation with current sampling resistor grounded.Supposing the equivalent circuit of the electrode-skinsurface with two elements, the voltage um(t) during thetime of low output of the monostable M.M.1 fallsexponentially according to the relation:

( )SXX RRCt

SXm e

RR

RUtu

+

+

= 0

0

00)( . (2)

-10

-8

-6

-4

-2

0

2

4

6

8

10

12

-0,04 -0,02 0 0,02 0,04

t[ms]

um[V]

Fig. 4. The measured voltage waveform when pulsation ismeasured with current sampling resistor grounded

In all of the shown measurement of the blood pulsationand respiration the current sampling resistor R0 is 682 and the voltage U0is 11 V.The sampling time Tis 1.34 s.

-0,5

-0,4

-0,3

-0,2

-0,1

0

0,1

0,2

0,3

0,4

0,5

0 0,5 1 1,5 2

t[s]

uout[V]

Fig. 5. The output voltage for blood pulsation in thearteries with grounded current sampling resistor

Fig. 5 shows blood pulsation measured in arteries with thecurrent sampling resistor grounded. Although the circuithas tuning ability of the output offset, the output voltagehas a variable DC component so it is difficult to adjust the

8/13/2019 Features of Voltage Pulse Plethysmography

3/6

DC component to zero. Because of it, a high pass filterCR is connected as is shown in Fig. 2.Fig. 6 shows the measured voltage um(t) during pulsationsmeasurement with one electrode grounded. Supposing theequivalent circuit of the electrode-skin interface with twoelements, during the low output of the monostable M.M.1the voltage um(t) rises exponentially according to:

( )

+= +

SXX RRC

t

SXm e

RRRUtu 0

000 1)( . (3)

-10

-8

-6

-4

-2

0

2

4

6

8

10

12

-0,04 -0,02 0 0,02 0,04 0,06

t[ms]

um[V]

Fig. 6. The measured voltage waveform when measuringpulsation with bio-tissue grounded

This measurements with bio-tissue resistance grounded ashis waveform shown on the Fig. 6 has essentially thesame exponential wave shape as it in current pulse

plethysmography and with similar characteristics.Although the initial supposition was that the measurementwith the electrode grounded will exhibit less noise, aftermultiple testing on several people, a significant advantageof either of the two methods hasn't been noticed. Thereason lies in relatively small measuring objectimpedance, that is, in the small resistance value of thecurrent sampling resistor. The capacitive coupling withthe interference source in that case generates small,actually, the same and negligible signal noise. As nosignificant differences in measuring with a groundedcurrent sampling resistor and a grounded electrode have

been shown, the respiration measurement has been carriedout only with current sampling resistor grounded. Fig. 7shows the output voltage for normal breathing and Fig. 8shows the output voltage for deep breathing.

-2,5

-2

-1,5

-1

-0,5

0

0,5

11,5

2

2,5

3

0 2 4 6 8 10

t[s]

uout[V]

Fig. 7. Output voltage for normal breathing respirationwith the current sampling resistor grounded

-2,5-2

-1,5

-1

-0,5

0

0,5

1

1,5

2

2,5

3

0 1 2 3 4 5 6 7 8 9 10

t[s]

uout[V]

Fig. 8. Output voltage for deep breathing respiration withthe current sampling resistor grounded

Determining the equivalent circuit of the

electrode-skin interface

When the supposed equivalent circuit of the electrode-skin interface consists of elements and the parameters forit are calculated, the approximate voltage waveform, that

is, the calculated voltage waveform based on theseparameters differs from the measured voltage. Butbecause the measuring sample has to be taken as soon aspossible, after starting moment in the time T(1 s 2 s)and S&H keep its value, the parallel capacitivecomponent CPXand resistive parallel componentRPXhaveno remarkable influence on measuring value.Fig. 9 shows the measurement circuit for determining the

parameters of the equivalent circuit electrode-skininterface.

Fig. 9. The measurement circuit for determining theparameters of the equivalent circuit electrode-skininterface

A simple analysis gives the expression for the voltageum(t):

PXSX

t

SX

PX

PXSXm

RRR

URe

RR

R

RRR

URtu

+++

+

++=

0

00

00

00 1)( , (4)

where isPXSX

XSXPX

RRR

CRRR

++

+=

0

01

)( . (5)

RSX and RPX are calculated by the end points method,that is, they are calculated from um(t) at the startingmoment (t=0+) and the end moment (distant enough

8/13/2019 Features of Voltage Pulse Plethysmography

4/6

from the starting moment, when the transient processactually dies out, t->). The expression for um(t) at thestarting moment is:

SXm

RR

URu

+=+

0

00)0( . (6)

The expression for um() is:

PXSX

m

RRR

URu

++

=0

00)( . (7)

Which gives:

+= 1

)0(0

0m

SXu

URR , (8)

+

=

)0(

1

)(

100

mmPX

uuURR . (9)

Inserting the expression (8) and (9) into (5) results in:

=++

+=

PXSX

XSXPX

RRR

CRRR

0

01

)(

000

000

000

)0(

1

)(

11

)0(1

)0()0(

1

)(

1

Ruu

Uu

U

CR

u

URR

uu

UR

mmm

X

mmm

+

+

++

+

+

+

= (10)

Further we get:

[ ]2001

)0(

)()0(

+

+=

m

mmX

u

uuCUR . (11)

Inserting the expression (8) and (9) into (4) results in:

[ ] )()()0()( 1 ++=

m

t

mmm ueuutu . (12)

Further follows:

)()0(

)()(1

+

=

mm

mm

t

uu

utue

, (13)

and expression for 1:

+

=

)()0(

)()(ln

1

mm

mm

uu

utu

t . (14)

For determining the CXvalue the point is taken when thevoltage umfalls to one half of its value:

2

)0()( 1

+= mm

uTu . (15)

Fig. 10 shows the measured voltage curve and the valuesare marked which are used to calculate the equivalent

circuit parameters of the electrode-skin interface.1is calculated as follows:

+

+

=

)()0(

)()0(5,0ln

11

mm

mm

uu

uu

T . (16)

XC is calculated from 1by formula:

SXX

RC 1

= . (17)

Fig. 10. The measured voltage curve with marked valuesused in calculation of the equivalent circuit parameters ofelectrode-skin interface

Fig. 11 shows the measured voltage (average value fromthree measurements) and the calculated value, which iscalculated inserting the values obtained by the end

points method and from the point when the voltage dropto one half of its value according to (8), (9) and (17). The

curve is given for the excitation voltage U0=11 V. Fig. 12shows the discrepancy of the calculated curve from themeasured one.

0

1

2

3

4

5

6

0 0,01 0,02 0,03 0,04t[ms]

um[V]

measured value calculated value

Fig. 11. Measured voltage compared to the calculatedvalue for U0=11 V, gel-wetted electrodes

-0,2

-0,1

0

0,1

0,2

0,3

0,4

0,5

0 0,005 0,01 0,015 0,02 0,025 0,03 0,035 0,04

t[ms]

Absoluteerror[V]

Fig. 12. Absolute error of the calculated curve forU0=11 V excitation

Since by determining the equivalent circuit parameters forthe electrode-skin interface with three elements a certain

8/13/2019 Features of Voltage Pulse Plethysmography

5/6

variance of the measured from the calculated voltagecurve is obtained, it was examined if determining the

parameters from some other three points would give lessvariance. Although the results of determination from threearbitrarily chosen points gave equally good results as theend-points method and the point when the voltage dropsto the half value, this method is considered more exact(the initial and the final voltage value can be well

determined because they don't depend on the capacitanceCX) than the choice of any three pints, so in the study thevalue parameters of the equivalent circuit were calculatedin this way.There is a noticeable variance of the measured value fromthe calculated value obtained by calculating from theequivalent circuit parameters of the electrode-skininterface determined by any of the procedures.Figures 13, 14 and 15 show the dependences of theresistanceRSX, the capacitance CXand the resistance RPXof the equivalent circuit electrode-skin interface withthree elements determined by the end-points methodand from the point when the voltage drops to its half

value. The dependence on the voltage U0 for the gel-wetted electrodes and for dry electrodes is shown.

0

200

400

600

800

1000

1200

0 2 4 6 8 10 12U0[V]

RSX[]

gel-wetted electrodes dry electrodes

Fig. 13. Dependence of the resistance RSXon the voltageU0

0

1

2

3

4

5

6

7

0 2 4 6 8 10 12U0[V]

CX[nF]

gel-wetted electrodes dry electrodes

Fig. 14. Dependence of the capacitance CXon the voltageU0

0

5

10

15

20

25

30

35

0 2 4 6 8 10 12U0[V]

RPX[k]

gel-wetted electrodes dry elecrodes

Fig. 15. Dependence of the resistance RPXon the voltageU0

When using gel-wetted electrodes the RSX values don'tvary considerably, the value range is between 710 and790 . When using dry electrodes there are also noconsiderable variations, the value range is between 800 and 1000 , that is, the values are somewhat higher thanwhen using gel-wetted electrodes (when the transient skin

surface resistance is lower).Neither when using gel-wetted electrodes nor when usingdry electrodes the capacitance CX values don't varysignificantly. When using gel-wetted electrodes thecapacitance value range is between 4 nF and 6 nF. Whenusing dry electrodes the value range is between 0.6 nF and1.5 nF, that is, the values are smaller than when using gel-wetted electrodes. CX is higher when using gel-wettedelectrodes than with dry electrodes because the tissuesoaked with gel acts as if the distance between theelectrodes has become shorter so the capacitance is larger.For the RPXdependence on U0a negative trend has beennoticed. When using the gel-wetted electrodes the values

vary between 13 k and 27 k. When using dryelectrodes the values vary between 13 kand 32 k.RPX(13 k 32 k) has no much influence on measurements,

because it is much higher thanRSX(700 - 1000 ).

Conclusion

Two voltage pulse plethysmography implementations arepossible with the current sampling resistor grounded andwith one electrode grounded. Blood pulsationsmeasurement in arteries and the respirations for bothimplementations have been carried out. Although it wassupposed that the implementation with one electrodegrounded will show less noise, this supposition has not

been confirmed. On the contrary, no significant differencein the quality of obtained measured results in the twoimplementations has been noticed.In the voltage pulse plethysmography the equivalentcircuit parameters of the electrode-skin interface withthree elements have been determined from the measuredvoltage characteristics. The absolute error was given as ameasure for the calculated voltage variance that iscalculated from the obtained equivalent circuit parametersand the measured voltage. A significant variance has beennoticed. That can be explained by impossibility of

8/13/2019 Features of Voltage Pulse Plethysmography

6/6

substituting the electrode-skin interface with a threelinear-element circuit and points out the need to modifythe electrode-skin interface with an equivalent circuit withnon-linear elements.But series resistance RSX is significant in thismeasurements while the capacitive component CPX and

parallel resistance RPX have very little influence whensampling time is short (1 s 2 s). Also it is important to

know, that maximal sensitivity can be obtained whensampling series resistance and bio-resistance are equal.Besides, it should be mentioned, that voltage pulsegenerator circuit is simpler than current pulse generatorcircuit.

References

[1] A. anti, T. tritof & V. Bilas, Plethysmographymeasurements using short current pulses with low duty-cycle, 20th Ann. Int. Conf. of the IEEE EMBS, HongKong, 1998, 1889-1892.[2] A. anti, D. Kovai& V. Bilas, Some new aspectsof electrical impedance end pulse plethysmography,

European Medical and Biological Engineering Conf.EMBEC, Vienna, Austria, 1999, 114-115.[3] W. Franks, W. Schenker, P. Schmutz & A.Hierlemann, Impedance characterization and modeling ofelectrodes for biomedical applications,IEEE Transactionson Biomedical Engineering52(7), 2005, 1295-1302.[4] L.A. Geddes, Electrodes and the measurement ofbioelectric events(Houston: John Wiley & Sons, 1972).[5] A.L. Klyuev, Z.A. Rotenbergb & V.V. Batrakov,Impedance of a passive iron electrode in a solutioncontaining a reducing agent, Russian Journal of

Electrochemistry41(1), 2005, 87-90.[6] A. anti, Pulse plethysmography in the blood

pressure measurement at the finger, 6th

IMEKO Conf. onMeasurement in Clinical Medicine, Sopron, 1990, 29-31.