SIGNAL SOURCES FOR SIMULATION, TESTING, AND CALIBRATION - Design and Development of Medical Electronic Instrumentation - page 297

inner surface. Their heart source model was a set of copper disks oriented along one of the
body axes and adjustable from outside the tank by means of a rod. The tank included
inhomogeneous regions constructed from cork and sandbags to simulate the electrical dis-
tortions caused by the spine, ribs, and lungs. Other investigators have used in fl ated dog

lungs and agar gel models of human lungs inserted into the tank.
If your research calls for an electrolytic body-shaped tank, a visit to your local hobby
and crafts store will prove invaluable. Describe what part of your anatomy you would like
to make a cast of, and the store's attendant's will help you select the best materials and
techniques. (Hint: don't ask for “michaplast.”) For the electrolyte solution you can use iso-
conductive saline made by mixing 1.8 g of noniodized salt (use Kosher salt if you don't
want to overpay for analytical-grade NaCl) per liter of distilled water (supermarket dis-
tilled water is okay). Electrodes should be Ag/AgCl, and inhomogeneous regions to simu-
late the lungs and other organs can be cast of high-density agarose gel (e.g., any standard
agarose gel with about 1200 g/cm 2 strength, melting range about 87 to 89°C, gelling range
about 36 to 39°C, sold by VWR Scienti c) or synthetic sponge foam.

As shown in Figure 6.33, a bipolar electrode to model cardiac potentials can be made from
two Ag/AgCl disk electrodes, some plastic tubing, and epoxy adhesive. This bipolar source is
fashioned after the one described by Brandon et al. [1971]. It yields a source with uniform dis-
tributions of current, 3 mm in diameter, separated by 12.7 mm. The potentiostat-galvanostat
of Figure 6.31 makes an excellent current source to drive this bipolar electrode.
Electrolyte tanks have also been used to investigate the way in which currents applied
through skin electrodes distribute within the body to cause tissue stimulation. This is an
important issue in external pacing, where high-voltage pacing pulses applied through skin-
surface electrodes not only cause the heart to contract, but cause quite a bit of pain to the
patient. Another area of special interest in the solution of the same problem of stimulation
over

fi

ow to nontarget tissue is functional neuromuscular stimulation (FNS; see Chapter 7).
We [Sagi-Dolev et al., 1995] used a phantom model in which a saline tank model was
improved by adding a layer to simulate skin impedance properties. The purpose of the
study was to look for electrode array geometries required to achieve target muscle activa-
tion with minimal over

fl

ow and to avoid pain or burning.
The tank was a small 40 cm

fl

filled with isoconductive
saline solution to simulate the volume conductor of the human forearm. To simulate the
way in which the skin's epidermis and panniculus adiposus distort the spatiotemporal char-
acteristics of FNS stimulation pulses, we made a carbon-loaded silicone elastomer that had
similar electrical properties to skin.
Rosell's method for the measurement of skin impedance [Rosell et al., 1988] was
adapted to measure the electrode-skin impedance for the speci

30 cm 30 cm glass aquarium fi

fi c conductive-silicone elec-

trode materials used for FNS and for the speci c spectral content of the FNS waveform

(1 to 9 kHz). A stimulating ring with inside diameter 12.5 mm and outside diameter 15 mm
was constructed from carbon-loaded silicone electrode material (Ag/AgCl electrodes
of similar dimensions were used by Rosell). Rectangular 2 mm

fi

5 mm voltage-sensing
electrodes were positioned, one within the center of the stimulating ring and the second
parallel to the

first at a distance of 2 mm from the ring as shown in Figure 6.34. The elec-
trodes were embedded at a depth of 1.5 mm within an isolating cast. The 1.5-mm cavities
allowed for a consistent conductive gel volume. A large self-adhesive Ag/AgCl ECG elec-
trode was used as a reference. The isolated ampli

fi

fi er input impedances had an equivalent

input impedance of 1 M in parallel with 20 pF.

The average impedance measurements were 7.7 k

Ω

at 10 kHz for a
1-cm 2 electrode surface area. By assuming that a parallel resistor-capacitor suitably models
the skin impedance, these impedance values result in an equivalent of a 7.64-k

Ω at 1 kHz and 4.8 k Ω

Ω resistor in

parallel to a 2479-pF capacitor for a 1-cm 2 electrode surface area. The arti cial “skin” was

fabricated to match these values. Varying amounts of colloidal graphite (MacDermid
PTF4150

fi

fl flexible graphite conductive paste) and silicone rubber (RTV No. 159) were