Implementation of A Low Cost Prototype for Electrical Impedance

Tomography based on the Integrated Circuit for Body Composition

Measurement AFE4300

ıctor Hugo Mosquera

, Adrian Arregui

, Ramon Brag

and Carlos Felipe Rengifo

Department of Electronic, University of Cauca, Street 5 # 4-70, Popay

an, Colombia

Department of Electronic Engineering, Technical University of Catalonia, Barcelona, Spain

Keywords:

Electrical Impedance Tomography, Image Reconstruction, Conductivity Distribution.

Abstract:

Electrical impedance tomography (EIT) is a technique of image reconstruction of the electrical conductivity

distribution in a tissue or region under observation. An electrical system for EIT comprises complex hardware

and software modules, which are designed for a speciﬁc application which requires that the system to be able

to detect conductivity variations within the study object. The Front-End for body composition measurement,

AFE4300 from Texas Instruments allows a minimal implementation of an electrical impedance tomography

system. It is the main device in the development of the EIT system presented in this paper, this device injects

the current signal and measures the tensions generated on the study region boundary by 8 electrodes, the

image reconstruction software was developed on the National Instruments platform Labview. The system

includes a microcontroller PIC16F886 to conﬁgure the 8 channels for the deﬁnition of the patterns of injection

and measurement of signals, also deﬁnes the current signal frequency and the bluetooth communication with

the computer for the image reconstruction. The developed system was validated by a planar resistive phantom

(CardiffEIT phantom), obtaining a stable voltage measurement every 50 ms per pair of electrodes, and a signal

to noise ratio (SNR) maximum of 71.8 dB, for a current signal of 50 kHz. Additionally, tests were carried out

in a saline tank with a concentration of 4 g/L, the developed system can simultaneously estimate the presence

of conductive and non-conductive disturbances into the tank.

1 INTRODUCTION

Patients suffering from urological disease or spinal

cord injury usually have difﬁculties perceiving blad-

der fullness and voiding due to neurological damage

or muscular atrophy. If these patients do not empty

their bladders on time, voiding dysfunction can result

in urinary tract infections and urinary reﬂux, which

may even lead to renal failure. The clinic process

for bladder emptying is done by inserting a catheter

into the bladder to drain urine, this method is invasive

and may cause urinary tract infection, besides that not

respect the micturition desire of patients. There are

techniques that apply the ultrasound and pressure sen-

sors for the bladder volume measurement for to as-

sist the bladder emptying, with the disadvantages of

the high noise and low precision of measurements (Li

et al., 2016).

The impedance distribution measurement is an-

other technique to measure the bladder volume with

the aim of assisting the process of emptying blad-

der. Electrical impedance tomography (EIT) is a non-

invasive technique that allows to get intra-thoracic im-

ages. The EIT systems are based on the injection of

currents and on the measurement of the resulting po-

tentials at the boundary, by means of electrodes. In

EIT applications on biological tissues, the currents

used are of sinusoidal nature, with amplitudes of a

few mA and frequencies ranging from 1 and 100 kHz.

Known the potentials and the currents at the object

boundary to be analyzed, a method of image recon-

struction is used to estimate the electrical conductiv-

ity distribution inside of the region (Harikumar et al.,

2013).

The EIT has numerous applications in the med-

ical ﬁeld, successfully entering in the monitoring of

intracranial hemorrhages or hematomas (Ayati et al.,

2015), cancer detection (Gao et al., 2014), study of

pelvic ﬂuid accumulation (Li et al., 2016), pulmonary

ventilation analysis (Bordes et al., 2016), blood pres-

sure measurement (Proenc¸a et al., 2016), among oth-

ers. The non-invasive and radiation-free character

Mosquera, V., Arregui, A., Bragós, R. and Rengifo, C.

Implementation of a Low Cost Prototype for Electrical Impedance Tomography based on the Integrated Circuit for Body Composition Measurement AFE4300.

DOI: 10.5220/0006554901210127

In Proceedings of the 11th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2018) - Volume 1: BIODEVICES, pages 121-127

ISBN: 978-989-758-277-6

121

of the EIT makes this technique a good alternative

for supporting the diagnosis and monitoring of med-

ical pathologies (Harikumar et al., 2013), (Islam and

Kiber, 2014).

Many studies have been advanced in the imple-

mentation of EIT systems, oriented to the detection

and monitoring of medical pathologies, these works

focus on the development of efﬁcient and portable

equipment, using electronic programmable devices.

Within the studies developed we highlight the use of

FPGA

s (Field programmable Gate array) or DSP

(Digital Signal Processors), devices that allow to de-

velop tomographic system capable of generating up

to 50 images per second, which has promoted the

use of EIT in problems with a high variation of the

conductivity per unit of time, for example the mon-

itoring of the blood pressure (Proenc¸a et al., 2016),

(Balleza-Ordaz et al., 2015), (Bordes et al., 2016).

For EIT systems oriented to medical applications with

a low temporal variability of its conductivity, the use

of microcontrollers presents good results as evidenced

in (Chitturi et al., 2014), (Fouchard et al., 2014) and

(Huang et al., 2016), with a lower cost compared to

systems developed with DSP

s and FPGA

s. Appli-

cations such as the bladder emptying and studies of

the cranial cavity and the bone system are ﬁelds in

which low-frequency EIT systems can be used in pro-

cesses of monitoring and pathologies detection (Li

et al., 2016), (Ateﬁ et al., 2016), (Ron et al., 2016).

The aim of this paper is to propose a new, low

cost, 8 channels EIT system for rapid prototyping, in-

tended for monitoring bladder emptying, process that

need a low quantity of images per second, based on

the body composition measurement device of Texas

Instruments AFE4300. The hardware structure of the

system is presented in section 2. The algorithm for

the reconstruction of conductivity distribution images

is described in section 3 and the experimental results

in a saline tank are presented in section 4.

2 HARDWARE STRUCTURE

EIT systems require the injection of a sinusoidal

current of both constant amplitude and frequency

and also the measurement of the potential difference

across the electrodes around the boundary of the ob-

ject under study. The values of the injected current

and the potentials measured on the electrodes are used

as the inputs to the reconstruction algorithm, produc-

ing images of the electrical conductivity distribution.

The system presented in this work generates a current

signal of 833 µA at 50 kHz, which is injected by ad-

jacent pairs of electrodes (Texas-Instruments, 2012).

The measure of the potentials is carried out by using

the adjacent electrode pairs method. These injection

and measurement patterns can be modiﬁed by conﬁg-

uring the registers of the AFE4300.

The EIT system consists of a mixed front-end

(AFE4300), which has 8 ports for current injection

and 8 ports for potentials measurement, also inte-

grates the direct digital synthesizer (DDS), voltage-

controlled current source (VCCS), voltage sensing,

quadrature demodulator or full-wave rectiﬁer and the

multiplexing stages for injection and measurement.

The LabVIEW platform of National Instruments is

used for the communication with the hardware and

for the implementation of the image reconstruction

algorithm. A microcontroller PIC16F886 connected

to Bluetooth module HC06 is used as the interface

between the PC running Labview and the AFE4300

based system. The block diagrams of the overall sys-

tem is depicted in ﬁgure 1. A photograph of the card

with the electronic components is presented in 2.

Figure 1: EIT system diagram.

Figure 2: EIT system implemented.

2.1 Front-End AFE4300

To measure the body impedance, the AFE4300 gener-

ates a sinusoidal signal by means of a DDS. The fre-

quency of this signal can be programmed from a 10-

bit record. The DDS output signal feeds a 6-bit DAC

whose refresh rate is 1 msps. The high-frequency

components of the DAC output signal are eliminated

BIODEVICES 2018 - 11th International Conference on Biomedical Electronics and Devices

122

by means of a low pass band second-order ﬁlter with

a cut-off frequency of 150 kHz. The DC component

of the signal delivered by the ﬁlter is removed by

means of a in series external capacitor. One of the

capacitor terminals is the input signal to a voltage-

current converter that is connected to a multiplexer

and then to the current injection ports (IOUTX). The

injection pattern is programmed through the conﬁgu-

ration register ISW MUX. Figure 3 shows the scheme

of body composition meter module of the AFE4300

described.

The voltage to current conversion is made through

the following equation:

i(t) =





(1)

Being R

= 1.5 kΩ ± 20% an internal element of

the AFE4300. Considering the minimal value of R

the maximal current generated is equal to the RMS

value of the voltage (1V

/(2

√

2) divided by 1.2 kΩ,

equivalent to 294.5 µA, which is below the maxi-

mum allowable current for a human being that is 500

µA (Master and Mark, 2012).

Figure 3: Front-End AFE4300 (Texas-Instruments, 2012).

The object to which the current signal is injected

generates a voltage difference between the measuring

electrodes, due to the internal conductivity distribu-

tion of the object. This voltage difference is sensed

by the AFE4300 through the voltage ports (VSEN-

SEx). Voltage measurements are performed using the

adjacent electrode method, which is deﬁned by the

programming register VSENSE MUX of AFE4300.

Voltage measurements are entered into a differential

ampliﬁer, which determines voltage variation due to

conductivity changes within the study object. This

variation can be obtained by full wave rectiﬁcation

or by quadrature (I/Q). The ﬁrst option provides

the magnitude of the impedance, while the second

one generates the real and the imaginary parts of the

impedance (Texas-Instruments, 2012). For this devel-

opment the ﬁrst measurement option is used, which

after performing the full wave rectiﬁcation, a low pass

ﬁlter is used to generate a DC signal proportional to

the impedance module |Z| (Equation (2)).

T /2

sin(ω

t + θ)dt =

= K

(2)

K is the proportionality constant due to calibra-

tion. For this study, the calibration ports of the device

are not used, in order to have the availability of the

8 channels for the injection and measurement signals.

The voltage values in the electrodes are sent to the

PIC16F886, to be later transmitted via Bluetooth to

the computer where the image reconstruction of the

conductivity distribution is carried out.

As can be demonstrated, the system is of easy im-

plementation and consists of a few number of elec-

tronic components, making this a compact and a

portable system. Other EIT systems implemented

with FPGA

s, DSP

s or microcontrollers, despite

their promising results, are systems that involve sev-

eral modules such as VCCS, Multiplexing, AD and

DA converters and their respective interfaces, which

makes these proposals, complex systems, such as

those presented in (Bera and Nagaraju, 2009), (Kha-

lighi et al., 2012) and (Wi et al., 2014).

The system designed is also low cost with a cost

of 65.60e. The prices of the devices AFE4300,

PIC16F886 and HC06 respectively are 4.58e, 2.09e

and 11.61e; passive elements such as resistors and

capacitors, in addition to electrode connectors, 2000

mA LiPo 3.7 V battery and Power Cell LiPo Charger,

have a price of 47.32e. Other systems based on

FPGA Virtex or Ciclon (Santos et al., 2016), (Khan

et al., 2015), (Shi et al., 2016) are more expensive as

these devices exceed 280e, for which our proposal a

more economical prototype.

3 IMAGES RECONSTRUCTION

The process of obtaining a conductivity distribution

within the object from the current and voltage mea-

surements on boundary is called the reconstruction

algorithm. The reconstruction algorithms solve a

mathematical problem that is nonlinear and ill-posed.

There are several reconstruction methods, such as

the absolute, dynamic and multifrequency (Santos

and Simini, 2013). This work uses the dynamic

method, which performs the reconstruction process

Implementation of a Low Cost Prototype for Electrical Impedance Tomography based on the Integrated Circuit for Body Composition

Measurement AFE4300

123

from changes in conductivity, due to changes in

the voltages in the electrodes, this technique is also

known as differential image.

Considering that the injection and measurement

are carried out in pairs of adjacent electrodes, it is pos-

sible to detect a voltage drop in any pair of electrodes.

The gray region of the ﬁgure 4 shows the area whose

changes in conductivity (δσ) generate changes over

adjacent electrodes. These changes are a function of

the measured voltages in different time or frequency

and are estimated by the equation (3).

Figure 4: Conductivity changes Detection.

δσ(x,y) = −

∑

i=1

Φ(u

),u

+ ∆t), ω

) (3)

The change of conductivity δσ by a pixel in a

given position (x,y) is estimated by the sum of the

voltage changes (Φ) produced in all the equipotential

regions, deﬁned by the measuring electrodes, where

that pixel belongs. For each injection combination

(ω

), one pixel will then belong only to one equipo-

tential region u

; For the implementation of the re-

construction algorithm is deﬁned a weighting func-

tion W (x, y,ω

), which is related to the conductivity

sensitivity changes within the study object, then the

mathematical model that describes the conductivity

distribution is:

δσ = −

∑

i=1

Φ (u

),u

+ ∆t) , ω

)W (x, y,ω

)

(4)

The function Φ being calculated as:

Φ(u

),u

+ ∆t), ω

) = ln(

+ ∆t)

)

≈

+ ∆t) −u

)

(5)

The reconstruction method represented by the

equations (4) and (5) is implemented in C++ and

compiled in a DLL (Dynamic Link Library), to later

be used in a LabVIEW application.

4 EXPERIMENTAL TESTS

To verify the performance of the proposed EIT sys-

tem the signal to noise ratio (SNR) is evaluated, this

parameter estimates the precision of measurement,

which quantiﬁed the repeatability of measurement un-

der unchanged conditions. For the analysis of the

SNR, a current signal of 50 kHz is injected into a 2D

resistive phantom (Cardiff EIT Phantom). The SNR

is calculated as the quotient between the mean and the

standard deviation of each of the 30 measured frames.

Different values of the delay between measures were

tested 10, 25, 50 and 100 ms, with objective of to

found the best characteristics the SNR. The results are

presented in the table 1, which indicates that delay of

50 ms per measurement presents the best characteris-

tics of SNR. This because of when switching a new

channel, there is a transient that adds dispersion to

the measurement if the delay between switching and

acquiring is shorter than 25 ms. The dispersion im-

proves if we wait for 50 ms, but it does not improve

and even grows if the delay is 100 ms or more because

there is a slow drift in the measured voltages that can

be observed at the signal inputs, that is the reason of

the dispersion increase for longer delays.

Table 1: SNR for different measurement times.

SNR (dB)

Maximum Average Minimum

10 ms 40.7 17.82 4.55

25 ms 74.73 28.57 9.74

50 ms 71.81 47.77 24.04

100 ms 56.04 45.7 33.95

Once the measurement time has been deﬁned (50

ms), tests are carried out in a tank with a saline

solution with conductive and non-conductive distur-

bances. The test for the image

s reconstruction be-

gins by determining the conductivity distribution of

the tank with the saline solution, with a concentration

of salt of 4g/L, in order to deﬁne the frame of ref-

erence for the dynamic reconstruction of subsequent

images with the presence of disturbances. Once the

reference frame is determined, conductive and non-

conductive materials are inserted into the tank to eval-

uate the potential change over the surface electrodes.

In Figure 5 it is possible to observe the conductiv-

ity distribution within the tank without disturbances,

BIODEVICES 2018 - 11th International Conference on Biomedical Electronics and Devices

124

in which small variations in the conductivity distribu-

tion can be observed.

Figure 5: Reconstruction the conductivity distribution of the

tank with saline solution.

Figure 6 shows the image reconstruction of the

tank with a non-conductive disturbance (peek bar).

From this Figure it can be observed that the object

causes a disturbance that is represented by blue color

due to a negative conductivity change of −0.9. The

same procedure is carried out for a conductive distur-

bance object which generates a change of conductiv-

ity of +0.9 in the reconstructed image (Figure 7).

Figure 6: Reconstruction of conductivity distribution with

non-conductive artefact.

Figure 7: Reconstruction of conductivity distribution with

conductive artefact.

Finally, the test is carried out with both a conduc-

tive and a non-conductive simultaneously, to evaluate

the capacity of the proposed system to detect various

types of disturbances. The ﬁgure 8 shows test result,

in which the image clearly evidence the two objects

of disturbance, discriminating between disturbances

generated by the conductive and non-conductive ma-

terials.

Figure 8: Reconstruction of conductivity distribution with

non-conductive and conductive artefacts.

The results do not show good accuracy between

the disturbance and the reconstructed image, which

can be improved with a method to minimizing elec-

trode interface impedance and a more elaborated im-

age reconstruction algorithm. On the other hand,

the system allows to detect changes in conductivity

within a saline environment which makes this sys-

tem a viable alternative to performing the study of

slow physiological processes such as bladder emp-

tying (Li et al., 2016), hematomas and hemorrhages

studies (Aristovich et al., 2016).

The proposed system presents advantages over

other developed devices because the AFE4300 con-

centrates the functions of: (i) electrical current gen-

eration, (ii) voltage measurement, (iii) multiplexing,

and (iv) demodulation, reducing the modules of the

system and its respective interfaces, facilitating its

implementation at a low cost. Designs like the one

presented in (Bera and Nagaraju, 2009), which em-

ploys a MAX038 for the generation of the current

signal and a QuadTech7600 for the potential sensing,

or in (Khalighi et al., 2012) that uses a XR2206 for

waveform generation and a CD4067B for multiplex-

ation, besides other modules, which makes the sys-

tem a complex alternative to implement. On the other

hand in (Wi et al., 2014) is presented a system called

Khu Mark 2.5, is a fairly complex modular equipment

that achieves 100 frames per second, but with a high

economic cost.

5 CONCLUSION

The designed system has a maximum SNR of 71.81

dB, which allows detecting conductivity variations in

a saline tank. The time delay of 50 ms between mea-

surements makes this prototype a good alternative for

Implementation of a Low Cost Prototype for Electrical Impedance Tomography based on the Integrated Circuit for Body Composition

Measurement AFE4300

125

the study of pathologies that do not require a high

frames frequency.

The proposed system requires a few electronic

components, which makes it easy to implement. On

the other hand its characteristics can be improved by

using more advanced methods of images reconstruc-

tion, which contribute to the decrease of the effects

of the noise and to have a better SNR (Hadinia and

Jafari, 2015), (Islam and Kiber, 2014).

REFERENCES

Aristovich, K. Y., Packham, B. C., Koo, H., dos Santos,

G. S., McEvoy, A., and Holder, D. S. (2016). Imag-

ing fast electrical activity in the brain with electrical

impedance tomography. NeuroImage, 124, Part A:204

– 213.

Ateﬁ, S. R., Seoane, F., Kamalian, S., Rosenthal, E.,

Lev, M., and Bonmassar, G. (2016). Intracranial

haemorrhage alters scalp potential distributions in

bioimpedance cerebral monitoring applications : pre-

liminary results from FEM simulation on a realistic

head model and human subjects. Medical Physics,

43(2):675–686. QC 20170111.

Ayati, S. B., Bouazza-Marouf, K., and Kerr, D. (2015).

In vitro localisation of intracranial haematoma using

electrical impedance tomography semi-array. Medi-

cal engineering & physics, 37(1):34–41.

Balleza-Ordaz, M., Perez-Alday, E., Vargas-Luna, M., and

Riu, J. (2015). Tidal volume monitoring by electrical

impedance tomography (EIT) using different regions

of interest (ROI): Calibration equations. Biomedical

Signal Processing and Control, 18:102 – 109.

Bera, T. K. and Nagaraju, J. (2009). A Simple instrumen-

tation calibration technique for Electrical Impedance

Tomography (EIT) using a 16-electrode phantom. In

2009 IEEE International Conference on Automation

Science and Engineering, pages 347–352.

Bordes, J., Goutorbe, P., Cungi, P. J., Boghossian, M. C.,

and Kaiser, E. (2016). Noninvasive ventilation during

spontaneous breathing anesthesia: an observational

study using electrical impedance tomography. Jour-

nal of Clinical Anesthesia, 34:420 – 426.

Chitturi, V., Farrukh, N., Thiruchelvam, V., and Fei, T. K.

(2014). A Low Cost Electrical Impedance Tomogra-

phy (EIT) for Pulmonary Disease Modelling and Di-

agnosis. In The Second International Conference on

Technological Advances in Electrical, Electronics and

Computer Engineering (TAEECE2014), pages 83–89.

The Society of Digital Information and Wireless Com-

munication.

Fouchard, A., Noca, A., Bonnet, S., Pham, P., Sinniger, V.,

Clarenon, D., and David, O. (2014). Modular archi-

tecture of a multi-frequency electrical impedance to-

mography system: Design and implementation. In

2014 36th Annual International Conference of the

IEEE Engineering in Medicine and Biology Society,

pages 6076–6079.

Gao, J., Yue, S., Chen, J., and Wang, H. (2014). Classiﬁca-

tion of normal and cancerous lung tissues by electrical

impendence tomography. Bio-medical materials and

engineering, 24(6):2229–2241.

Hadinia, M. and Jafari, R. (2015). An element-free galerkin

forward solver for the complete-electrode model in

electrical impedance tomography. Flow Measurement

and Instrumentation, 45:68–74.

Harikumar, R., Prabu, R., and Raghavan, S. (2013). Electri-

cal impedance tomography (EIT) and its medical ap-

plications: a review. Int J Soft Comp Eng, 3(4):193–8.

Huang, J.-J., Hung, Y.-H., Wang, J.-J., and Lin, B.-S.

(2016). Design of wearable and wireless electrical

impedance tomography system. Measurement, 78:9–

17.

Islam, M. R. and Kiber, M. A. (2014). Electrical impedance

tomography imaging using gauss-newton algorithm.

In Informatics, Electronics & Vision (ICIEV), 2014 In-

ternational Conference on, pages 1–4. IEEE.

Khalighi, M., Vahdat, B. V., Mortazavi, M., Hy, W., and

Soleimani, M. (2012). Practical design of low-cost in-

strumentation for industrial Electrical Impedance To-

mography (EIT). In Instrumentation and Measure-

ment Technology Conference (I2MTC), 2012 IEEE In-

ternational, pages 1259–1263. IEEE.

Khan, S., Manwaring, P., Borsic, A., and Halter, R. (2015).

FPGA-based voltage and current dual drive system

for high frame rate electrical impedance tomography.

IEEE Transactions on Medical Imaging, 34(4):888–

901.

Li, R., Gao, J., Li, Y., Wu, J., Zhao, Z., and Liu, Y. (2016).

Preliminary study of assessing bladder urinary volume

using electrical impedance tomography. Journal of

Medical and Biological Engineering, 36(1):71–79.

Master, T. and Mark, T. (2012). Medical electrical equip-

ment part 1: General requirements for basic safety and

essential performance.

Proenc¸a, M., Braun, F., Muntan

e, E., Sol

a, J., Adler, A.,

Lemay, M., Thiran, J.-P., and Rimoldi, S. F. (2016).

Non-invasive monitoring of pulmonary artery pres-

sure at the bedside. In Engineering in Medicine and

Biology Society (EMBC), 2016 IEEE 38th Annual

International Conference of the, pages 4236–4239.

IEEE.

Proenc¸a, M., Braun, F., Sol

a, J., Thiran, J.-P., and Lemay,

M. (2016). Noninvasive pulmonary artery pressure

monitoring by EIT: a model-based feasibility study.

Medical & Biological Engineering & Computing,

pages 1–15.

Ron, A., Abboud, S., and Arad, M. (2016). Home mon-

itoring of bone density in the wrista parametric EIT

computer modeling study. Biomedical Physics & En-

gineering Express, 2(3):035002.

Santos, E. and Simini, F. (2013). Comparison of electri-

cal impedance tomography reconstruction techniques

applied to IMPETOM system. In Bioinformatics and

Bioengineering (BIBE), 2013 IEEE 13th International

Conference on, pages 1–4. IEEE.

Santos, S. A., Robens, A., Boehm, A., Leonhardt, S., and

Teichmann, D. (2016). System Description and First

BIODEVICES 2018 - 11th International Conference on Biomedical Electronics and Devices

126

Application of an FPGA-Based Simultaneous Multi-

Frequency Electrical Impedance Tomography. Sen-

sors (14248220), 16(8):1 – 21.

Shi, X., You, F., Xu, C., Ji, Z., Liu, R., Dong, X., Fu,

F., and Huo, X. (2016). Design and implementa-

tion of a high-precision electrical impedance tomog-

raphy data acquisition system for brain imaging. In

2016 IEEE Biomedical Circuits and Systems Confer-

ence (BioCAS), pages 9–15.

Texas-Instruments (2012). Datasheet-AFE4300-Low-Cost,

Integrated Analog Front-End for Weight-Scale and

Body Composition Measurement. Dallas, Texas,

USA.

Wi, H., Sohal, H., McEwan, A. L., Woo, E. J., and Oh,

T. I. (2014). Multi-frequency electrical impedance to-

mography system with automatic self-calibration for

long-term monitoring. IEEE transactions on biomed-

ical circuits and systems, 8(1):119–128.

Implementation of a Low Cost Prototype for Electrical Impedance Tomography based on the Integrated Circuit for Body Composition

Measurement AFE4300

127