Professional Documents
Culture Documents
discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/23269750
CITATIONS
DOWNLOADS
VIEWS
10
65
169
6 AUTHORS, INCLUDING:
Kevin R Aroom
Charles S Cox
SEE PROFILE
SEE PROFILE
Brijesh S Gill
University of Texas Health Science Center
35 PUBLICATIONS 377 CITATIONS
SEE PROFILE
Background. Bioimpedance analysis has found utility in many fields of medical research, yet instrumentation can be expensive and/or complicated to build.
Advancements in electronic component design and
equipment allow for simple bioimpedance analysis using equipment now commonly found in an engineering
lab, combined with a few components exclusive to impedance analysis.
Materials and methods. A modified Howland bridge
circuit was designed on a small circuit board with connections for power and bioimpedance probes. A programmable function generator and an oscilloscope were
connected to a laptop computer and were tasked to
drive and receive data from the circuit. The software
then parsed the received data and inserted it into a
spreadsheet for subsequent data analysis. The circuit
was validated by testing its current output over a range
of frequencies and comparing measured values of impedance across a test circuit to expected values.
Results. The system was validated over frequencies
between 1 and 100 kHz. Maximum fluctuation in current was on the order of micro-Amperes. Similarly, the
measured value of impedance in a test circuit followed
the pattern of actual impedance over the range of
frequencies measured.
Conclusions. Contemporary generation electronic
measurement equipment provides adequate levels of
connectivity and programmability to rapidly measure
and record data for bioimpedance research. These
components allow for the rapid development of a simple but accurate bioimpedance measurement system
To whom correspondence and reprint requests should be addressed at Department of Surgery, University of TexasHouston,
6431 Fannin St., MSB 4.268, Houston, TX 77030. E-mail: brijesh.s.
gill@uth.tmc.edu.
23
0022-4804/09 $36.00
2009 Elsevier Inc. All rights reserved.
24
Z cos() and Z sin(), respectively, where Z is the magnitude of the impedance and is the phase angle. Fig.
2 gives a graphical representation of impedance and its
components.
Measurement of electrical impedance takes advantage
of the relationship between impedance, voltage, and current. Ohms law states the relation V IZ, where V is
voltage, I is current, and Z is impedance. By injecting a
controlled amount of current into a section of tissue, the
resulting voltage across that tissue provides an easily
acquired signal for recording and subsequent analysis.
Alternating current (AC) is used as the source of electrical current because it prevents iontophoresis and allows
determination of the phase angle shift, a property that
cannot be measured if direct current (DC) is used. Modern oscilloscopes have the ability to automatically measure phase angle differences between 2 signals. Knowl-
FIG. 3. Block diagram describing major components of the bioimpedance measurement system. A Printed Circuit Board (PCB)
contains the current source and connections for the electrode probes.
Attached to the PCB are a DC power supply, function generator, and
an oscilloscope, which are controlled and monitored by a program
running on a laptop PC. Data from the experiment is automatically
uploaded to the laptop throughout the measurement.
25
TABLE 1
TABLE 2
Components of system
Component
INA128P Instrumentation
Amplifier
LF412 CN Dual Amplifier
0.1 F X7R Ceramic
Capacitor
10 k Metal Film
Resistor 1%
51 k Metal Film
Resistor 1%
Manufacturer
Number
Texas Instruments
National Semiconductor
Vishay
1
3
Digi-Key
Digi-Key
FIG. 4. Schematic of the modified Howland Voltage Controlled Current Source (VCCS). An instrumentation amplifier is used as a
preamplifier to the signal entering the circuit from an external function generator. The signal then enters the actual Howland bridge, which
consists of 2 operational amplifiers and a few resistors and capacitors. The output of the circuit is connected to the electrode pair that will
actually measure tissue impedance through an oscilloscope probe that measures the voltage across the 2 electrodes. This schematic and
the PCB layout file are available online at http://www.uth.tmc.edu/pediresearch under howlandbridge.sch and edemameter.v123,
respectively.
26
Input/Output Hardware
To drive and record the response of the circuit, both a function
generator and an oscilloscope were controlled by software on a laptop
computer to inject current at various frequencies. The software allows for customization of the waveform and also makes data collection and storage very simple by exporting data directly to standard
spreadsheet format. A function generator (33220A; Agilent Technologies, Santa Clara, CA) was used to provide the input voltage signal
to the VCCS. A mixed signal oscilloscope (MSO6032A; Agilent Technologies) measured and recorded voltage and phase angle data from
the signal across the 2 electrodes. These devices are accurate both in
generating a signal as well as in measuring voltage, with 300-MHz
bandwidth and 12 bits of resolution on the oscilloscope. High bandwidth and resolution allows for the recording of smooth waveforms.
The large input impedance of the oscilloscope probe minimizes the
influence of the measuring device as a current pathway.
Software Programming
Both the oscilloscope and the function generator were connected to
a laptop computer (Dell Latitude, 1 GHz CPU, 256 MB RAM, Windows XP) via USB cables. The control software (VEE Pro 8.0, Agilent
Technologies) was loaded onto the computer. This software allows
the user to control supported electronic equipment attached to the
host computer, allowing duplex communication for both sending
commands and receiving data. The software also communicates with
spreadsheet software (Microsoft Excel) to automatically generate
spreadsheets with the data measured. Programs are developed in an
intuitive graphical format using lines of data and commands connecting different module and subroutine boxes. Fig. 5 shows how
the program is structured within VEE. A front panel graphical user
interface is also able to be generated to allow for user input during
execution of the program and is shown in Fig. 6. Copies of the VEE
software code used for this design are available online at
www.uth.tmc.edu/pediresearch under brainimpedanceprogram.vee.
FIG. 5. Layout of the VEE program used to control hardware. The program consists of graphical modules that represent commands, data
structures such as arrays, and elements that make up the graphical user interface (GUI). Within this environment, the user can customize
the frequencies over which impedance measurements are taken, as well as other attributes such as the duration of sampling and the order
in which samples are taken. The file containing this program can be found at the following URL: http://www.uth.tmc.edu/pediresearch under
the filename brainimpedanceprogram.vee. (Color version of figure is available online.)
27
Surface mount resistors (0.1%) were tested to determine if using high tolerance resistors in the VCCS
could improve the profile of current throughout the
range of relevant frequencies. Fig. 9 shows the comparison of the no-load current profiles for both the 1%
tolerance resistors and the 0.1% tolerance resistors.
Minimal improvement can be seen in that there is still
fluctuation in the amount of injected current outside
the bounds of 1 and 100 kHz.
FIG. 6. The front panel of the program where the user selects the
test method and test subject. The front panel is designed to simplify
the measurement process, resulting in faster and more efficient data
collection. Feedback on the frequency being used and the resulting
RMS value can be shown on this screen if preferred.
RESULTS
Calibration
DISCUSSION
28
FIG. 7. No-Load current response of the constant current driver circuit. A multimeter was placed in series between the 2 electrodes and
measured the current produced over a range of frequencies from 150 Hz to 500 kHz. The current varies from 49 A at 100 Hz to 15 A at
500 kHz. However, current remains rather constant between 1 kHz and 100 kHz. (Color version of figure is available online.)
FIG. 8. Measured voltage of test circuit versus expected values. Expected values were derived from the measured values of individual
resistance and capacitance elements in the test circuit and applied to the governing equations of Ohms Law to determine the frequency based
impedance. The measured impedance was simply the measured RMS voltage divided by the current injected into the tissue by the circuit.
The value of injected current was referenced from the corresponding value shown in Fig. 7. Deviation is relatively large from 100 Hz to 10
kHz but corresponds well between 10 and 500 kHz. (Color version of figure is available online.)
29
FIG. 9. Comparison of the no-load current response between 1% tolerance resistors and 0.1% tolerance resistors shows a minimal
difference. The variations present below 1 kHz and above 100 kHz are present for both types of resistors and could be attributed to the
performance of the operational amplifiers. Matched values of resistors are an important requirement of this circuit design; so the use of high
precision resistors on the order of 1% or less tolerance is highly recommended. (Color version of figure is available online.)
ison of impedance can still be performed using 2 electrodes, but the drift associated with electrode polarization reduces the overall accuracy. Another attribute of
the design that requires consideration, depending on the
application, is the choice of electrode. The platinum electrodes used for in vivo brain tissue measurements may
not be the best choice for impedance measurements
taken in other tissues or surfaces. For example, skin
surface bioimpedance analysis would use conventional
Ag/AgCl gel electrodes or electrode bands that wrap
around the circumference of a limb.
In any case, this system can be easily adapted to a
probe construction that includes 4 or more electrodes,
eliminating the polarization source of measurement
error. If the drive circuitry is connected to the lateral
pair of electrodes, while voltage is measured across the
medial pair, then a tetra-polar configuration has been
established.
As with any device that is made from primary components, it is very important to perform calibration and
verification of this system to ensure constant current
delivery and to generate a calibration curve which
could be used in determining impedance values after
data collection has been performed.
By incorporating equipment that is widely available in
research laboratories, parts exclusively used for the purpose of bioimpedance analysis can be reduced to the PCB
and its components, costing less than $50. Creating the
layout for a PCB is very simple using free software such
as Express PCB (http://www.expresspcb.com/) or EagleCAD
(http://www.cadsoft.de/). The file present at http://uth.tmc.
30
3.
4.
Bernstein DP, Lemmens HJ. Stroke volume equation for impedance cardiography. Med Biol Eng Comput 2005;43:443.
5.
6.
7.
Radhakrishnan RS, Shah K, Xue H, et al. Measurement of intestinal edema using an impedance analyzer circuit. J Surg Res
2007;138:106.
8.
ACKNOWLEDGMENTS
This work was supported by NIH Grant T32 GM008792-06
(M.T.H.) and TATRC Grant W81XWH-07-1-0496 (B.S.G./C.S.C.).
REFERENCES
1.
2.