Bioimpedance principles

interpretation on a better knowledge of the immittivity of
BIOIMPEDANCE
the smaller tissue components (2–11).
SVERRE GRIMNES
ØRJAN G. MARTINSEN 1. TYPICAL BIOIMPEDANCE DATA
University of Oslo
Oslo, Norway
Figure 1 shows the three most common electrode systems.
With two electrodes, the current carrying electrodes and
signal pick-up electrodes are the same (Fig. 1, left). If the
Bioimpedance describes the passive electrical properties electrodes are equal, it is called a bipolar lead, in contrast
of biological materials and serves as an indirect transdu- to a monopolar lead. With 3-(tetrapolar) or 4-(quadropo-
cing mechanism for physiological events, often in cases lar) electrode systems, separate current carrying and
where no specific transducer for that event exists. It is an signal pick-up electrodes exist. The impedance is then
elegantly simple technique that requires only the applica- transfer impedance (12): The signal is not picked up from
tion of two or more electrodes. According to Geddes and the sites of current application.
Baker (1), the impedance between the electrodes may The 4-electrode system (Fig. 1, right) has separate pick-
reflect ‘‘seasonal variations, blood flow, cardiac activity, up (PU) and current carrying (CC) electrodes. With ideal
respired volume, bladder, blood and kidney volumes, voltage amplifiers, the PU electrodes are not current
uterine contractions, nervous activity, the galvanic skin carrying, and therefore, their polarization impedances do
reflex, the volume of blood cells, clotting, blood pressure not introduce any voltage drop disturbing measured tissue
and salivation.’’ impedance. In the 3-electrode system (Fig. 1, middle), the
Impedance Z [ohm, O] is a general term related to the measuring electrode M is both a CC and signal PU
ability to oppose ac current flow, expressed as the ratio electrode.
between an ac sinusoidal voltage and an ac sinusoidal
current in an electric circuit. Impedance is a complex 1.1. A 4-Electrode Impedance Spectrum
quantity because a biomaterial, in addition to opposing
current flow, phase-shifts the voltage with respect to the Figure 2 shows a typical transfer impedance spectrum
current in the time-domain. Admittance Y [siemens, S] is (Bode plot) obtained with the 4-electrode system of Fig. 1
the inverse of impedance (Y ¼ 1/Z). The common term for (right). It shows two dispersions (to be explained later).
impedance and admittance is immittance (2). The transfer impedance is related to, but not solely
The conductivity of the body is ionic (electrolytic), determined by, the arm segment between the PU electro-
because of for instance Na þ and Cl– in the body liquids. des. As we shall see, the spectrum is determined by the
The ionic current flow is quite different from the electronic sensitivity field of the 4-electrode system as a whole. The
conduction found in metals: Ionic current is accompanied larger the spacing between the electrodes, the more the
by substance flow. This transport of substance leads to results are determined by deeper tissue volumes. Even if
concentrational changes in the liquid: locally near the all the electrodes are skin surface electrodes, the spectrum
electrodes (electrode polarization), and in a closed-tissue is, in principle, not influenced by skin impedance or
volume during prolonged dc current flow. electrode polarization impedance.
The studied biomaterial may be living tissue, dead For many, it is a surprise that the immittance mea-
tissue, or organic material related to any living organism sured will be the same if the CC and PU electrodes are
such as a human, animal, cell, microbe, or plant. In this interchanged (the reciprocity theorem).
chapter, we will limit our description to human body
tissue. 1.2. A 3-Electrode Impedance Spectrum
Tissue is composed of cells with poorly conducting,
Figure 3 shows a typical impedance spectrum obtained
thin-cell membranes; therefore, tissue has capacitive
with three skin surface electrodes on the underarm (Fig.
properties: the higher the frequency, the lower the im-
pedance. Bioimpedance is frequency-dependent, and im-
pedance spectroscopy, hence, gives important information
about tissue and membrane structures as well as intra-
and extracellular liquid distributions. As a result of these
M1
capacitive properties, tissue may also be regarded as a M PU1
dielectric (3). Emphasis is then shifted to ac permittivity M2 PU PU2
(e) and ac losses. In linear systems, the description by CC1
permittivity or immittivity contains the same information.
It must also be realized that permittivity and immittivity
are material constants, whereas immittance is the directly
CC CC2
measured quantity dependent on tissue and electrode
geometries. In a heterogeneous biomaterial, it is impos- Figure 1. Three skin surface electrode systems on an underarm.
sible to go directly from a measured immittance spectrum Functions: M ¼ measuring and current carrying, CC ¼ current
to the immittivity distribution in the material. An impor- carrying, PU ¼ signal pick-up.
tant challenge in the bioimpedance area is to base data
1

Wiley Encyclopedia of Biomedical Engineering, Copyright & 2006 John Wiley & Sons, Inc.

2 BIOIMPEDANCE

1, middle). Notice the much higher impedance levels than high frequencies (4 100 kHz), the results are dominated
found with the 4-electrode system. The measured zone is by deeper layer tissues.
under M and comprises electrode polarization, skin im- Figure 3 also shows the effect of contact electrolyte
pedance, and deeper layer impedance, all physically in penetration into the initially dry skin (three curves: at the
series. The electrode polarization impedance is a source of moment of electrode onset on dry skin, after 1 h, and after
error; electrode impedance is not tissue impedance. At low 4 h). The electrode polarization contribution can be judged
frequencies (o 1000 Hz), the result is dominated by the by studying Fig. 11. Initially, the electrode polarization
high impedance of the human skin with negligible influ- impedance has negligible influence on the LF results;
ence from the polarization impedance of the electrode. At however, at HF, around 20% of the measured impedance
is from the M electrode itself.
Also, the immittance measured will be the same if the
102
CC and PU electrodes are interchanged (the reciprocity
theorem).
|Z|

2. FROM MAXWELL TO BIOIMPEDANCE EQUATIONS

101 The Maxwell equation most relevant to bioimpedance is:
100 101 102 103 104 105 106
Frequency (Hz) r Â H À @D=@t ¼ J ð1Þ
5
D ¼ eo E þ P ð2Þ
0
−5 where H ¼ magnetic field strength [A/m], D ¼ electric flux
theta

−10 density [coulomb/m2], J ¼ current density [A/m2], E ¼
−15 electric field strength [V/m], eo ¼ permittivity of vacuum
[farad (F) /m], and P ¼ electric polarization, dipole mo-
−20
100 101 102 103 104 105 106 ment pr. volume [coulomb/m2].
If the magnetic component is ignored, Equation 1 is
Frequency (Hz)
reduced to:
Figure 2. Typical impedance spectrum obtained with four equal
electrodes attached to the skin of the underarm as shown on Fig. 1
@D=@t ¼ À J ð3Þ
(right). All electrodes are pregelled ECG-electrodes with skin gel-
wetted area 3 cm2. Distance between electrode centers: 4 cm.
Equations 1–3 are extremely robust and also valid under
nonhomogeneous, nonlinear, and anisotropic conditions.
They relate the time and space derivatives at a point to
the current density at that point.
106
020219d.z60
Impedance and permittivity in their simplest forms are
105 020219k.z60 based on a basic capacitor model (Fig. 4) and the introduc-
020219y.z60
tion of some restrictions:
104
|Z|

a) Use of sufficiently small voltage amplitude v across
103 the material so the system is linear. b) Use of sinusoidal
functions so that with complex notation a derivative (e.g.,
102 @E/@t) is simply the product joE (j is the imaginary unit
100 101 102 103 104 105 106
and o the angular frequency). c) Use of D ¼ eE (space
Frequency (Hz)
vectors), where the permittivity e ¼ ereo, which implies
that D, P, and E all have the same direction, and there-
−15 fore, that the dielectric is considered isotropic. d) No fringe
effects in the capacitor model of Fig. 4.
−40
theta

−65

−90
100 101 102 103 104 105 106
Frequency (Hz)
Figure 3. Typical impedance spectrum obtained with the 3-
electrode system on the underarm as shown on Fig. 1 (middle).
The parameters are: at the time of electrode onset on dry skin,
after 1 h, and after 4 h of contact. Figure 4. The basic capacitor model.

BIOIMPEDANCE 3

Under these conditions, a lossy dielectric can be char- scalar. Tissue, as a rule, is anisotropic. Plonsey and Barr
0 00
acterized by a complex dielectric constant e ¼ e À je or a (5) discussed some important complications posed by
complex conductivity r ¼ s þ js [S/m], then s ¼ oe00 (2).
0 00 0
tissue anisotropy and also emphasized the necessity of
Let us apply Equation 3 on the capacitor model where the introducing the concept of the bidomain. A bidomain
metal area is A and the dielectric thickness is L. However, model is useful for cardiac tissue, where the cells are
Equation 3 is in differential form, and the interface connected by two different types of junctions: tight junc-
between the metal and the dielectric represents a discon- tions and gap junctions where the interiors of the cells are
tinuity. Gauss law as an integral form must therefore be directly connected. The intracellular space is one domain
used, and we imagine a thin volume straddling an area of and the interstitial space the other domain.
the interface. According to Gauss law, the outward flux of
D from this volume is equal to the enclosed free charge
density on the surface of the metal. With an applied 3. GEOMETRY, SENSITIVITY AND RECIPROCITY
voltage v, it can be shown that D ¼ ve/L. By using Equation
3, we then have @D/@t ¼ jove/L ¼ J, and i ¼ joveA/L ¼ vjoC. Resistivity r [O Á m] and conductivity s [S/m] are material
We now leave the dielectric and take a look at the constants and can be extended to their complex analogues:
external circuit where the current i and the voltage v impedivity [O Á m] and admittivity [S/m]. The resistance of
(time vectors) are measured and the immittance deter- a cylinder volume with length L, cross-sectional area A,
mined. The admittance is Y ¼i/v [siemens, S]. With no and uniform resistivity r is:
losses in the capacitor, i and v will be phase-shifted by 901
(the quadrature part). The conductance is G ¼ s0 A/L [S], R ¼ rL=A: ð6Þ
and the basic equation of bioimpedance is then (time
vectors): Equation 6 shows how bioimpedance can be used for
volume measurements (plethysmography). Notice, how-
Y ¼ G þ joC ð4Þ ever, that, for example, a resistance increase can be
caused either by an increased tissue length, a reduced
Three important points must be made here: cross-sectional area, or an increased resistivity. Tissue
First, Equation 4 shows that the basic impedance dielectric and immittivity data are listed by Duck (13).
model actually is an admittance model. The conductive Figure 5 illustrates typical resistance values for body
and capacitive (quadrature) parts are physically in paral- segments (2), valid without skin contribution and without
lel in the model of Fig. 4. current constrictional effects caused by small electrodes.
Second, the model of Fig. 4 is predominantly a dielectric By using the term ‘‘resistance,’’ we indicate that they are
model with dry samples. In bioimpedance theory, the not very frequency-dependent. Notice the low resistance of
materials are considered to be wet, with double layer
and polarization effects at the metal surfaces. Errors are
introduced, which, however, can be reduced by introdu-
cing 3- or 4-electrode systems (Fig. 1). Accordingly, in
dielectric theory, the dielectric is considered as an insu-
lator with dielectric losses; in bioimpedance theory, the
material is considered as a conductor with capacitive
properties. Dry samples can easily be measured with a 100
2-electrode system. Wet, ionic samples are prone to errors 110
and special precautions must be taken.
Third, Equations 1–3 are valid at a point. With a 13
homogeneous and isotropic material in Fig. 4, they have
the same values all over the sample. With inhomogeneous 270
and anisotropic materials, the capacitor model implies
values averaged over the volume. Then, under linear
50
(small signal) conditions, Equation 4 is still correct, but
the measured values are difficult to interpret. The capa- 500
citor is basically an in vitro model with a biomaterial
values
placed in the measuring chamber. The average anisotropy
can be measured by repositioning the sample in the in ohm
capacitor. In vivo measurements, as shown in Fig. 1, 140
must be analyzed from sensitivity fields, as shown in the
next chapter.
From Equation 3, the following relationship is easily 320
deduced (space vectors):

J ¼ sE: ð5Þ
Figure 5. Typical body segment resistance values. From (2), by
permission.
Equation 5 is not valid in anisotropic materials if s is a

4 BIOIMPEDANCE

the thorax (13 O) and the high resistance of one finger (500 7 in the same way, and the interchange of the PU and CC
O). electrodes do not change the value of R. Equation 7 is
Figure 6 shows the effect of a constrictional zone caused therefore based on the reciprocity theorem (14).
by one small electrode. Such an electrode system is called In a monopolar or bipolar electrode system, Equation 7
monopolar because most of the measuring results are a simplifies to
result of the impedance of the small electrode proximity
zone.
Z
R¼ rJ 2 dv S ¼ J2; ð9Þ
3.1. Sensitivity Field of an Electrode System V

Intuitively, it is easy to believe that if a small tissue
volume changes immittivity, the influence on the mea-
where J [1/m2] is the local current density caused by a
surement result is larger the nearer that tissue volume is
unity current passed through the electrode pair.
to the electrodes, which is indeed the case and is illu-
The analysis of the sensitivity field of an electrode
strated by an equation based on the work of Geselowitz
system is of vital importance for the interpretation of
(14):
measured immittance. Figure 7 (top) illustrates, for in-
Z stance, the effect of electrode dimensions in a bipolar
R¼ rJ CC Á J reci dv; ð7Þ electrode system. As the gap between the electrodes
V
narrows, the local sensitivity in the gap increases accord-
ing to S ¼ J2. However, the volume of the gap zone also
where R is the transfer resistance [O] measured by a 4- becomes smaller, and the overall contribution to the
electrode system; r is the local resistivity in the small
integral of Equation 7 is not necessarily dominating. If
volume dv (if r is complex, the integral is the transfer local changes (e.g., from a pulsating blood artery) is to be
impedance Z); JCC [1/m2] is the local current density space picked up, the artery should be placed in such a high-
vector in dv caused by the CC electrodes carrying a unity
sensitivity zone.
current; and Jreci [1/m2] is the local current density space Figure 7 (bottom) illustrates the effect of electrode-
vector in dv caused by the PU electrodes if they also electrode center distance in a bipolar system. As distance
carried a unity current (reciprocal excitation).
is increased, the sensitivity in deeper layers will increase
The product JCC Á Jreci is the local dot vector product, but still be small. However, large volumes in the deeper
which may be called the local sensitivity S [1/m4] of the layers will then have a noticeable effect, as small volumes
electrode system:
proximal to the electrodes also have.
S ¼ J CC Á J reci : ð8Þ

Unity current is used so that sensitivity is a purely
geometrical parameter not dependent on any actual cur-
rent level. S is a scalar with positive or negative values in
different parts of the tissue; the spatial distribution of S is
the sensitivity field.
The implications of Equation 7 are important, and at
first sight counter intuitive: In a 4-electrode system, both
electrode pairs determine the sensitivity, not just the PU
electrodes as one may intuitively believe. There will be
zones of negative sensitivity in the tissue volume between
the PU and CC electrodes. The zones will be dependent on, Figure 7. The measuring depth as a function of electrode
for example, the distance between the PU and CC electro- dimensions (top) and electrode spacing (bottom). From (2), by
des. The PU and CC current density fields enter Equation permission.

constrictional
zone

segmental zone:
R = ρ L/A
segmental
zone Figure 6. Left: Monopolar system with one electrode
much smaller than the other. Increased resistance
from the constrictional current zone with increased
current density. Right: Segment resistance with uni-
form current density, large bipolar electrodes. From
(2), by permission.

BIOIMPEDANCE 5

4. ELECTRICAL MODELS prisingly effective addition is also to replace the capacitor
C by a more general CPE (Constant Phase Element). A
Bioimpedance is a measure of the passive properties of CPE is not a physical device but a mathematical model,
tissue, and, as a starting point, we state that tissue has you cannot buy a CPE as you buy a resistor (j ¼ 01) or a
resistive and capacitive properties showing relaxation, capacitor (j ¼ 901). A CPE can have any constant phase
but not resonance, phenomena. As shown by Equation 4, angle value between 01 and 901, and mathematically, it is
admittance is a very simple device (2). Figure 8 shows a popular
equivalent circuit in two variants.
Y ¼ G þ joCp One such circuit deﬁnes one dispersion (15), character-
ized by two levels at HF and LF, with a transition zone
j ¼ arc tanðoCp =GÞ where the impedance is complex. Both at HF (Z ¼ RN) and
at LF (Z ¼ R þ 1/Gvar), the impedance Z is purely resistive,
jYj2 ¼ G2 þ ðoCp Þ2 ; ð10Þ determined by the two ideal resistors. The circuit of Fig. 8
(left) is with three ideal, frequency-independent compo-
nents, often referred to as the Debye case, and the im-
where j is the phase angle indicating to what extent the
pedance is:
voltage is time-delayed, G is the parallel conductance [S],
and Cp the parallel capacitance [F].
1
The term oCp is the capacitive susceptance. Z ¼ R1 þ t ¼ C=Gvar :
Impedance is the inverse of admittance (Z ¼ 1/Y); the Gvar þ Gvar jot
equations are:
The diagram to the right in Fig. 8 is with the same two
ideal resistors, but the capacitor has been replaced by a
Z ¼ R À j=oCs
CPE (2). The equivalent circuit of a CPE consists of a
resistor and a capacitor, both frequency-dependent so that
j ¼ arc tanðÀ1=oRCs Þ
the phase becomes frequency-independent.
jZj2 ¼ R2 þ ð1=oCs Þ2 : ð11Þ 1
Z ¼ R1 þ
Gvar þ G1 ðjotÞa
The term À 1/oCs is the capacitive reactance.
The values of the series (Rs, Cs) components values are ja ¼ cosðap=2Þ þ j sinðap=2Þ ð13Þ
not equal to the parallel (1/G, Cp) values:
In Equation 13, the CPE admittance is G1(jot)a; and t may
Z ¼ Rs À j=oCs ¼ G=jYj2 À joCp =jYj2 : ð12Þ be regarded as a mean time constant of a tissue volume
with a distribution of different local time constants. t may
Equation 12 illustrates the serious problem of choosing, also be regarded just as a frequency scaling factor; ot is
for example, an impedance model if the components are dimensionless and G1 is the admittance value at the
physically in parallel: Rs and Cs are both frequency- characteristic angular frequency when ot ¼ 1. a is related
dependent when G and Cp are not. Implicit in these both to the constant phase j of the CPE according to ja and
equations is the notion that impedance is a series circuit j ¼ a Á 901 and to the frequency exponent in the term oa.
of a resistor and a capacitor, and admittance is a parallel This double inﬂuence of a presupposes that the system is
circuit of a resistor and a capacitor. Measurement results Fricke-compatible. According to Fricke’s law, the phase
must be given according to one of these models. A model angle j and the frequency exponent m are related in many
must be chosen, no computer system should make that electrolytic systems so that j ¼ m Á 901. In such cases, m is
choice. An important basis for a good choice of model is replaced by a (2,16).
deep knowledge about the system to be modeled. An A less general version of Equation 13 was given by Cole
electrical model is an electric circuit constituting a sub- (17):
stitute for the real system under investigation, as an
equivalent circuit. DR 1
Z ¼ R1 þ ¼ R1 þ : ð14Þ
One ideal resistor and one ideal capacitor can represent 1 þ ðjotÞa DG þ DGðjotÞa
the measuring results on one frequency, but can hardly be
expected to mimic the whole immittance spectrum actu- The Cole model does not have an independent conduc-
ally found with tissue. Usually, a second resistor is added tance Gvar in parallel with the CPE; DG controls both the
to the equivalent circuit, and one simple and often sur- parallel ideal conductance and the CPE, which implies

Gvar Gvar
R∞ R∞

C
CPE
Figure 8. Two versions of a popular equiva-
lent circuit.

6 BIOIMPEDANCE

that the characteristic frequency is independent of DG, in filled with electrolyte species. A double layer will be
the same way that t ¼ C/Gvar in Equation 11 would be formed in the electrolyte at the electrode surface. This
constant if both C and Gvar varied with the same factor. double layer represents an energy barrier with capacitive
The lack of an independent conductance variable limits properties. Both processes will contribute to electrode
the application of the Cole equation (18). All of Equations polarization immittance. Figure 11 shows the impedance
12–14 are represented by circular arc loci if the impedance spectrum of a commercial, pregelled ECG electrode of the
is plotted in the complex plane, but one of them must be type used for obtaining the results in Figs. 2 and 3.
chosen.
Instead of Bode plots, a plot in the complex Argand or
Wessel (2) plane may be of value for the interpretation of
the results. Figure 9 shows the data of Fig. 2 as a Z and a 5.1. Electrode Designs
Y plot. The two dispersions are clearly seen as more or less Skin surface electrodes are usually made with a certain
perfect circular arcs. The LF dispersion is called the a- distance between the metal part and the skin [Fig. 12
dispersion, and the HF dispersion is called the b-disper- (top)]. The enclosed volume is filled with contact electro-
sion (15). b-dispersion is caused by cell membranes and lyte, often in the form of a gel contained in a sponge. The
can be modeled with the Maxwell–Wagner structural stronger the contact electrolyte, the more rapid the pene-
polarization theory (15,19). The origin of the a-dispersion tration into the skin. There are two surface areas of
is more unclear. concern in a skin surface electrode: The area of contact
Figure 8 is a model well-suited for skin impedance: The between the metal and the electrolyte determines the
RN is the deeper tissue series resistance and the parallel polarization impedance; the electrolyte wetted area of
combination represents the stratum corneum with inde- the skin (the effective electrode area, EEA) determines
pendent sweat duct conductance in parallel. Figure 10 the skin impedance.
shows another popular model better suited for living Figure 12 (bottom) shows needle electrodes for invasive
tissue and cell suspensions. The parallel conductance G0 measurements. Some types are insulated out to the tip;
is the extracellular liquid, the capacitance is the cell others have a shining metal contact along the needle
membranes, and the R is the intracellular contributions. shaft. Some are of a coaxial type with a thin center lead
isolated from the metal shaft.

5. ELECTRODES AND INSTRUMENTATION

The electrode is the site of charge carrier transfer, from G0
electrons to ions or vice versa (20–22). The electrode
proper is the contact zone between the electrode metal
(electronic conduction) and the electrolyte (ionic conduc- C
R
tion). As ionic current implies transport of substance, the
electrolytic zone near the metal surface may be depleted or Figure 10. Tissue or suspension equivalent circuit.

0.05 75
011005a.z60 011005a.z60

0.04

50
0.03

Real Center: 0,017258
Imag. Center: −0,0042139
Y''

Diameter: 0,018898 Real Center: 77,829
0.02 25 Imag. Center: 22,034
Deviation: 0,00028145
Z''

Low Intercept: 0,0088009 Diameter: 90,173
High Intercept: 0,025715 Deviation: 1,074
Depression Angle: −26,485 Low Intercept: 38,493
0.01 High Intercept: 117,16
Depression Angle: 29,255
1MHz 1kHz w_max: 50,495
0
LF
0 1Hz
1Hz 1kHz
1MHz LF

−0.01 −25
0 0.01 0.02 0.03 0.04 0.05 0.06 0 25 50 75 100
Y' Z'

Figure 9. The data in Fig. 2 shown in the complex Wessel diagram as Y and Z plots. Circular arc
fits for the same LF dispersion are shown in both plots.

BIOIMPEDANCE 7

signs have extended it below 10 Hz. Now, lock-in ampli-
103
020604h.z60
fiers are the preferred instrumentation for bioimpedance
measurements (2). The lock-in amplifier needs a synchro-
nizing signal from the signal oscillator for its internal
102
|Z|

synchronous rectifier. The output of the lock-in amplifier
is a signal not only dependent on input signal amplitude,
but also phase. A two-channel lock-in amplifier has two
101 rectifiers synchronous with the in-phase and quadrature
100 101 102 103 104 105 106 oscillator signals. With such an instrument, it is possible
Frequency (Hz) to measure complex immittance directly. Examples of
20
commercially available instruments are the SR model
series from Stanford Research Systems, the HP4194A,
10 and the Solartron 1260/1294 system with an extended LF
0 coverage (10 mHz–32 MHz). Bioimpedance is measured
theta

−10
with small currents so that the system is linear: An
applied sine waveform current results in a sine pick-up
−20 signal. A stimulating electrode, on the other hand, is used
−30 with large currents in the nonlinear region (7), and the
100 101 102 103 104 105 106 impedance concept for such systems must be used with
Frequency (Hz) care.
Figure 11. Electrode polarization impedance, one pregelled ECG Software for Bode plots and complex plane analysis are
commercial electrode of the type used in Figs. 2 and 3. The commercially available; one example is the Scribner
positive phase at HF is caused by the self-inductance of the ZView package. This package is well-suited for circular
electrode wire. arc fits and equivalent circuit analysis.

5.3. Safety
insulator: metal: In a 2- and 3-electrode system, it is usually possible to
EA(cm2)
gel:
operate with current levels in the microampere range,
corresponding to applied voltages around 10 mV rms. For
{

most applications and direct cardiac, these levels may be
EEA(cm2) safe (24,25).
With 4-electrode systems, the measured voltage for a
given current is smaller, and for a given signal-to-noise
ratio, the current must be higher, often in the lower mA
range. For measuring frequencies below 10 kHz, this is
unacceptable for direct cardiac applications. LF mA cur-
metal shaft
insulated shaft

rents may also result in current perception by neuromus-
cular excitation in the skin or deeper tissue. Dependent on
the current path, these LF current levels are not necessa-
rily dangerous, but are unacceptable for routine applica-
tions all the same.

6. SELECTED APPLICATIONS

uni- bi- coax bipolar rings 6.1. Laboratory-on-a-Chip
polar + refer.
With microelectrodes in a small sample volume of a cell
Figure 12. Top: skin surface types. Bottom: invasive needles.
suspension, it is possible to manipulate, select, and char-
From (2), by permission.
acterize cells by rotation, translocation, and pearl chain
formation (26). Some of these processes are monitored by
bioimpedance measurements.
With modern technology, it is easy to fabricate micro-
electrodes with small dots or strips with dimensions in the
micrometer range that are well-suited for single-cell dis- 6.2. Cell Micromotion Detection
crimination. A monopolar microelectrode is convenient to study cell
attachment to a surface. Many cell types need an attach-
ment to flourish, and it can be shown that measured
5.2. Instrumentation and Software
impedance is more dominated by the electrode surface
Bridges achieve high precision. The frequency range is the smaller it is. Cell micromotion can be followed with nm
limited (23), although modern self-balanced bridge de- resolution on the electrode surface (27).

8 BIOIMPEDANCE

6.3. Cell Suspensions 6.7. Skin Moisture
The Coulter counter counts single cells and is used in The impedance of the stratum corneum is dependent on its
hospitals all over the world. The principle is based on a cell water content. By measuring the skin susceptance, it is
suspension where the cells and the liquid have different possible to avoid the disturbance of the sweat duct parallel
impedivities. The suspension is made to flow through a conductance (31), and hence assess the hydration state of
capillary, and the capillary impedance is measured. In the stratum corneum. Low-excitation frequency is needed
addition to rapid cell counting, it is possible to character- to avoid contribution from deeper, viable skin layers.
ize each cell on passing (28).
6.8. Skin Fingerprint
6.4. Body Composition Electronic fingerprint systems will, in the near future,
Bioimpedance is dependent on the morphology and im- eliminate the need for keys, pincodes, and access cards in
pedivity of the organs, and with large electrode distances, a number of daily-life products. With a microelectrode
it is possible to measure segment or total body water, matrix or array, it is possible to map the fingerprint
extra- and intracellular fluid balance, muscle mass, and electrically with high precision using bioimpedance mea-
fat mass. Application areas are as diversified as sports surements. Live finger detection is also feasible to ensure
medicine, nutritional assessment, and fluid balance in that the system is not fooled by a fake finger model or a
renal dialysis and transplantation. Body composition in- dead finger (32).
struments represent a growing market. Many of them are
single-frequency instruments, and with the electrode sys- 6.9. Impedance Tomography
tems used, it is necessary to analyze what they actually
By applying many electrodes on the surface of a body, it is
measure (29).
possible to map the distribution of immittivity in the
volume under the electrodes (33–35). One approach is to
6.5. Impedance Plethysmography use, for example, 16 electrodes, excite one pair and
arrange a multichannel measurement of the transfer
See that entry in this encyclopedia. impedance in all the other unexcited pairs (Fig. 13). By
letting all pairs be excited in succession, one complete
measurement is performed. By choosing a high measuring
6.6. Impedance Cardiography (ICG) and Cardiac Output
frequency of, for example, 50 kHz, it is possible to sample
A tetrapolar system is used with two band electrodes data for one complete image in less than one tenth of a
around the neck, one band electrode corresponding to second, and live images are possible. The sensitivity in
the apex of the heart, and the fourth further in caudal Equation 9 clearly shows that it is more difficult to obtain
direction A more practical system uses eight spot electro- sharp spatial resolution the larger the depth from the skin
des arranged as four double electrodes, each with one PU surface. In practice, the resolution is on the order of
and one CC electrode. The amplitude of the impedance centimeters; therefore, other advantages are pursued
change DZ as a function of the heartbeat is about 0.5% of (e.g., the instrumentation robustness and the simplicity
the baseline value. The DZ waveform is similar to the of the sensors).
aorta blood pressure curve. The first time derivative dZ/dt
is called the impedance cardiographic curve (ICG). By
6.10. Monitoring Tissue Ischemia and Death
adding information about patient age, sex, and weight, it
is possible to estimate the heart stroke volume and cardiac Large changes in tissue impedance occur during ischemia
output. The resistivity of blood is flow-dependent (11), and (36,37), tissue death, and the first hours afterward (2).
as long as the origin (heart-, aorta-, lung-filling/emptying) The changes are related to changed distribution of intra-
of the signal is unclear, the cardiac output transducing cellular and extracellular liquids, variations in the gap
mechanism will also be obscure. Sensitivity field analysis junctions between the cells, and, in the end, the break-
may improve this status (30). down of membrane structures.

i i

1 2 3 1 2 3
16 4 16 4
Better-
15 5 15 5 conducting
object
6 6
14 14
7 7
13 v 13 v
8 8
12 12
9 9 Figure 13. Principle of a tomography setup.
11 10 11 10 From (2), by permission.

BIOIMPEDANCE 9

7. FUTURE TRENDS 17. K. S. Cole, Permeability and impermeability of cell mem-
branes for ions. Cold Spring Harbor Symp. Quant. Biol. 1940;
Basic scientific topics: immittivity of the smaller tissue 8:110–122.
components, sensitivity field theory, impedance spectro- 18. S. Grimnes and Ø. G. Martinsen, Cole electrical impedance
scopy and tissue characterization, nonlinear tissue prop- model—a critique and an alternative. IEEE Trans. Biomed.
erties, and single-cell manipulation. Eng. 2005; 52(1):132–135.
Instrumentation: ASIC (Application Specific Integrated 19. H. Fricke, The Maxwell–Wagner dispersion in a suspension of
Circuit) design as a new basis for small and low-cost ellipsoids. J. Phys. Chem. 1953; 57:934–937.
instrumentation, also for single-use applications. Teleme- 20. J. R. Macdonald, Impedance Spectroscopy, Emphasizing Solid
try technology to improve signal pick-up and reduce noise Materials and Systems. New York: John Wiley, 1987.
and influence from common-mode signals. 21. J. Koryta, Ions, Electrodes and Membranes. New York: John
Applications: microelectrode technology; single-cell Wiley, 1991.
and microbe monitoring; electroporation; electrokinetics 22. L. A. Geddes, Electrodes and the Measurement of Bioelectric
(e.g., electrorotation); tissue characterization; monitoring Events. New York: Wiley-Interscience, 1972.
of tissue ablation, tissue/organ state, and death/rejection 23. H. P. Schwan, Determination of biological impedances. In: W.
processes. L. Nastuk, ed., Physical Techniques in Biological Research,
Vol. 6. San Diego, CA: Academic Press, 1963, pp. 323–406.
24. IEC-60601, Medical electrical equipment. General require-
ments for safety. International Standard, 1988.
BIBLIOGRAPHY
25. C. Polk and E. Postow, eds., Handbook of Biological Effects of
1. L. A. Geddes and L. E. Baker, Principles of Applied Biomedi- Electromagnetic Fields. Boca Raton, FL: CRC Press, 1995.
cal Instrumentation. New York: John Wiley, 1989. 26. R. Pethig, J. P. H. Burt, A. Parton, N. Rizvi, M. S. Talary, and
2. S. Grimnes and Ø. G. Martinsen, Bioimpedance and Bioelec- J. A. Tame, Development of biofactory-on-a-chip technology
tricity Basics. San Diego, CA: Academic Press, 2000. using excimer laser micromachining. J. Micromech. Microeng.
1998; 8:57–63.
3. R. Pethig, Dielectric and Electronic Properties of Biological
Materials. New York: John Wiley, 1979. 27. I. Giaever and C. R. Keese, A morphological biosensor for
mammalian cells. Nature 1993; 366:591–592.
4. S. Ollmar, Methods for information extraction from impe-
dance spectra of biological tissue, in particular skin and oral 28. V. Kachel, Electrical resistance pulse sizing: Coulter sizing.
mucosa—a critical review and suggestions for the future. In: M. R. Melamed, T. Lindmo, and M. L. Mendelsohn, eds.,
Bioelectrochem. Bioenerg. 1998; 45:157–160. Flow Cytometry and Sorting. New York: Wiley-Liss Inc., 1990,
pp. 45–80.
5. R. Plonsey and R. C. Barr, A critique of impedance measure-
ments in cardiac tissue. Ann. Biomed. Eng. 1986; 14:307–322. 29. K. R. Foster and H. C. Lukaski, Whole-body impedance—
what does it measure? Am. J. Clin. Nutr. 1996; 64(suppl):
6. J. Malmivuo and R. Plonsey, Bioelectromagnetism. Oxford,
388S–396S.
UK: Oxford University Press, 1995.
30. P. K. Kaupinen, J. A. Hyttinen, and J. A. Malmivuo, Sensi-
7. R. Plonsey and R. C. Barr, Bioelectricity. A Quantitative tivity distributions of impedance cardiography using band
Approach. New York: Plenum Press, 2000. and spot electrodes analysed by a 3-D computer model. Ann.
8. K. R. Foster and H. P. Schwan, Dielectric properties of tissues Biomed. Eng. 1998; 26:694–702.
and biological materials: a critical review. CRC Crit. Rev.
31. Ø. G. Martinsen and S. Grimnes, On using single frequency
Biomed. Eng. 1989; 17(1):25–104. electrical measurments for skin hydration assessment. Innov.
9. J.-P. Morucci, M. E. Valentinuzzi, B. Rigaud, C. J. Felice, N. Techn. Biol. Med. 1998; 19:395–399.
Chauveau, and P.-M. Marsili, Bioelectrical impedance tech- 32. Ø. G. Martinsen, S. Grimnes, K. H. Riisnæs, and J. B.
niques in medicine. Crit. Rev. Biomed. Eng. 1996; 24:223–681.
Nysæther, Life detection in an electronic fingerprint system.
10. S. Takashima, Electrical Properties of Biopolymers and Mem- IFMBE Proc. 2002; 3:100–101.
branes. New York: Adam Hilger, 1989. 33. K. Boone, Imaging with electricity: report of the European
11. K. Sakamoto and H. Kanai, Electrical characteristics of concerted action on impedance tomography. J. Med. Eng.
flowing blood. IEEE Trans. BME 1979; 26:686–695. Tech. 1997; 21:6,201–232.
12. F. F. Kuo, Network Analysis and Synthesis, international 34. P. Riu, J. Rosell, A. Lozano, and R. Pallas-Areny, Multi-
edition, New York: John Wiley, 1962. frequency static imaging in electrical impedance tomography.
13. F. A. Duck, Physical Properties of Tissue. A Comprehensive Part 1: instrumentation requirements. Med. Biol. Eng. Com-
Reference Book. San Diego, CA: Academic Press, 1990. put. 1995; 33:784–792.
14. D. B. Geselowitz, An application of electrocardiographic lead 35. J. G. Webster, Electrical Impedance Tomography. New York:
theory to impedance plethysmography. IEEE Trans. Biomed. Adam Hilger, 1990.
Eng. 1971; 18:38–41. 36. M. Schaefer, W. Gross, J. Ackemann, M. Mory, and M. M.
15. H. P. Schwan, Electrical properties of tissue and cell suspen- Gebhard, Monitoring of physiological processes in ischemic
sions. In: J. H. Lawrence and C. A. Tobias, eds., Advances in heart muscle by using a new tissue model for description of
Biological and Medical Physics, vol. V. San Diego, CA: Aca- the passive electrical impedance spectra. Proc. 11th Int. Conf.
demic Press, 1957, pp. 147–209. on Electrc. Bioimpedance, Oslo, Norway, 2001: 55–58.
16. H. Fricke, Theory of electrolytic polarisation. Phil. Mag. 1932; 37. M. Osypka and E. Gersing, Tissue impedance and the appro-
14:310–318. priate frequencies for EIT. Physiol. Meas. 1995; 16:A49–A55.

Bioimpedance principles

More Related Content

What's hot (20)

Viewers also liked (6)

Similar to Bioimpedance principles (20)

More from ES-Teck India (20)

Recently uploaded (20)

Bioimpedance principles