Flexible Array of Active Concentric Ring Electrodes for Surface

Bioelectrical Recordings

Application to Non-invasive Recordings of EEnG

J. Garcia-Casado

, V. Zena

, G. Prats-Boluda

, Y. Ye-Lin

and E. Garcia-Breijo

Grupo de Bioelectrónica, I3BH-UPV, Camino de Vera SN, Valencia, 46022, Spain

Centro de Reconoc. Molec. y Desarrollo Tecnológico. Ud Mixta UPV-UV,Camino de Vera SN, Valencia, 46022, Spain

Keywords: Ring Electrode, Laplacian Recording, Non-invasive Myoelecrical Recording, Electroenterogram.

Abstract: The estimation of laplacian potential on the body surface obtained by means of concentric ring electrodes

can provide bioelectrical signals with better spatial resolution and less affected by bioelectrical interferences

than monopolar and bipolar recordings with conventional disc electrodes. These ringed electrodes are

usually implemented on rigid substrates which cannot adapt to body surface curvature. In this paper an array

of flexible concentric ring electrodes for non-invasive bioelectrical activity recordings is presented. The

array contains three tripolar electrodes in bipolar configuration (TCB, inner disc and external ring are

shorted) which is suitable for body surface mapping. A preconditioning circuit module is directly connected

to the electrode array to perform a first stage of filtering and amplification as close as possible to the

recording area. Simultaneous recordings of intestinal myoelectrical activity (electroenterogram, EEnG) by

means of the flexible array of ringed-electrodes and by disc electrodes with gel were carried out in healthy

volunteers in fast state. Signals from the developed array of electrodes presented lower electrocardiographic

and respiratory interference than conventional bipolar recordings with disc electrodes. The small bowel’s

slow wave myoelectrical activity can be identified more easily in the ringed-electrodes recordings.

1 INTRODUCTION

1.1 Bioelectrical Laplacian Recordings

Surface recordings of bioelectrical signals are

usually recorded by means of disc electrodes in

bipolar or unipolar configuration. In the first method

the potential difference between a pair of electrodes

is measured. In the latter method the potential of

each electrode is compared either to a neutral

electrode or to the average of several electrodes. One

drawback of using conventional disc electrodes in

bioelectrical surface recordings is their poor spatial

resolution which is mainly caused by the blurring

effect of the different conductivities of the volume

conductor (Bradshaw et al., 2001). In this respect,

Laplacian has been shown to reduce the smoothing

effects caused by the volume conductor and to

increase the spatial resolution in localizing and

differentiating multiple dipole sources (Wu et al.,

1999); (Besio et al., 2006).

There are different approaches to estimate the

laplacian potential on the body surface. The first

ones to be used were discretization techniques like

the one introduced by Hjorth as early as in 1975

(Hjorth, 1975). In that study, the laplacian of the

EEG signal was estimated as the difference between

the average potential of four disc electrodes in the

form of a cross and the potential of a fifth disc

electrode placed in the center of the cross. In the late

80s, analytic solutions to estimate the laplacian of

the surface potential were proposed in order to

reduce discretization errors (Perrin et al., 1987).

These are complex discrete computational

techniques, generally not suitable for real-time

applications. Nevertheless, laplacian potential can

also be directly estimated by means of concentric

ring electrodes in tripolar, bipolar or tripolar in

bipolar configuration (TCB, where the outer ring

and the center disc were electrically shorted) (Lu

and Tarjan, 1999); (Besio et al., 2006); (Koka and

Besio, 2007).

TCB ring electrodes have already been used to

estimate the Laplacian potential of bioelectrical

signals such as the electrocardiogram (ECG),

Garcia-Casado J., Zena V., Prats-Boluda G., Ye-Lin Y. and Garcia-Breijo E..

Flexible Array of Active Concentric Ring Electrodes for Surface Bioelectrical Recordings - Application to Non-invasive Recordings of EEnG.

DOI: 10.5220/0004196500210027

In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2013), pages 21-27

ISBN: 978-989-8565-34-1

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

electroencephalogram (EEG), electroenterogram

(EEnG) and the electrohysterogram (EHG), so as to

increase the spatial resolution of and the signal

quality of conventional surface potential recordings

(Li et al., 2005); (Prats-Boluda et al., 2011); (Koka

and Besio, 2007). These electrodes are usually active

since the signals sensed by concentric ring

electrodes in laplacian configuration are weaker than

the ones obtained my conventional monopolar or

bipolar recordings, and the output impedance is

bigger. Nevertheless, the ringed-electrodes used in

these studies were developed on rigid substrates like

printed-circuit boards, what can provoke a poor

skin-to-electrode contact and discomfort to the

patient.

Therefore, the aim of this study is to develop

concentric ring electrodes on a flexible substrate to

join the advantages of laplacian recordings with the

comfort and better adaptation to the body surface

curvature of conventional disposable disc electrodes.

1.2 Intestinal Myoelectrical Activity

The study of intestinal motility is an outstanding

field in gastroenterology due to the fact that

abnormal motility patterns are related with several

intestinal pathologies (Quigley, 1996). This is the

case in irritable bowel syndrome, intestinal

obstruction, paralytic ileus, and bowel ischemia. The

main problem in monitoring intestinal activity is the

difficult anatomic access, hence most methods of

studying this activity are considered to be invasive.

One possible solution would be the recording of

intestinal myoelectrical activity on abdominal

surface. This signal is named Electroenterogram

(EEnG) and it is composed of two waves: slow

waves (SW) and spike burst (SB). The former are

periodical, omnipresent electrical potentials that

regulate the maximum rate of intestinal muscle

contractions. The latter are fast action potentials

which are located in the plateau of the SW. They are

only present when contractions appear. Whereas SW

are related to the frequency and propagation velocity

of the contractions (Weisbrodt, 1987), SB determine

the presence and the intensity of the contractions.

The frequency of the SW changes along the small

intestine from about 12cpm at duodenum to

8cpm at ileum (Fleckenstein and Oigaard, 1978)

There are few studies about abdominal surface

recordings to identify the EEnG in humans (Chen et

al., 1993); (Chang et al., 2006); (Prats-Boluda et al.,

2007); (Prats-Boluda et al., 2011). The main reason

is that human EEnG is a very weak signal, which is

severely attenuated especially in the SB frequency

range, because of the insulation effects of the

abdominal layers and spatial filtering (Garcia-

Casado et al., 2006). Surface EEnG is also very

sensitive to physiological interferences such as ECG

and respiration, being difficult to identify the SW

component of the EEnG by visual inspection of

abdominal surface recordings. The ECG spectral

frequency range overlaps the SB frequency range,

therefore it is necessary to eliminate it from

abdominal recordings to identify the SB component

of the EEnG (Garcia-Casado et al., 2006). As

regards to respiration interference, the typical

breathing frequency range (12cpm to 24cpm) is very

close to the frequency of the SW (8cpm to 12cpm),

so it is not possible to use conventional filters to

remove this interference.

Laplacian recordings of the EEnG by means of

active concentric ring electrodes on rigid substrates

have proven to enhance signal quality in comparison

to conventional monopolar and bipolar recordings

with disc electrodes. Therefore, a second objective

of this work is to test and study the possible benefits

of the flexible concentric ring electrodes to be

developed in this study, on the surface recordings of

the EEnG.

2 MATERIAL AND METHODS

2.1 Active Electrode Array Design and

Implementation

2.1.1 Sensing Part

In this work it has been decided to design an array of

electrodes rather than an individual electrode for

surface bioelectrical recordings, since this kind of

recordings are usually multichannel and moreover

laplacian recordings are often used for body surface

mappings. Specifically an array of three concentric

ring TCB active electrodes was developed. The

sensor is made out of two parts: a disposable sensing

part with three TCB electrodes and a reusable

battery-powered signal conditioning circuit. Each of

the three sensing electrodes consist of an inner disc

and two concentric rings in bipolar configuration i.e.

the disc and the outer ring are shorted together

(TCB).

The outer diameter of the external ring was set to

24mm which is a compromise between bigger

electrodes that would yield signals of higher

amplitude and smaller electrodes that would provide

better spatial resolution. The rest of dimensions of

the electrode are designed considering the following

BIODEVICES2013-InternationalConferenceonBiomedicalElectronicsandDevices

criteria:

- The sum of the areas of the outer ring (Aout) and

the inner disc (Ain) should be equal to the area of

the middle ring (Amid) so as to provide similar input

impedances, improving the common mode rejection

ratio.

- The distance between the inner disc and the

middle ring should be the same as the distance

between the middle and outer rings to reduce

common mode interferences.

Other issues such as a minimum recording area of

50mm

and the limitations of the implementation

technique were considered. The dimensions of each

ringed-electrode of the array are shown in figure 1.

As it can also be appreciated in this figure, an

opened-ring version has been designed in order to

avoid shorts in the layout of the traces from the

electrodes to the connectors. The flexible electrode

array was implemented by screen-printing

technology on polymer substrates. Specifically, a

biocompatible silver paste was printed (Dupont 5064

Silver conductor, thickness 17 μm) on Polyester

Melinex ST506 substrate (thickness 175μm). The

serigraphy was made by using an AUREL 900 High

precision screen stencil printer. Cured period of inks

was 130ºC for 10 minutes.

Figure 1: Dimensions (mm) of the concentric ring

electrode to be implemented in the array.

2.1.2 Signal Conditioning Part

As stated before, signals from concentric ring

electrodes are of very low amplitude, especially in

the cases that the bioelectrical signal to be recorded

on the body surface is weak. Therefore it is highly

recommended to include an amplification stage as

close as possible to the sensing electrode.

In this work a battery-powered conditioning

circuit was developed and directly connected to the

electrode array. Precisely, a 12 bias (only six are

used) flexible-flat-cable-to-flexible-printed-circuit

connector (TE Connectivity/AMP-1-84953-2

FFC/FPC) was used for the connection. The circuit

is composed of a preamplifier (gain 31.9), followed

by a coupling circuit (high pass cut off frequency

0.05Hz) and an additional differential amplification

stage (gain 106.1) for each of the three TCB

electrodes of the array. Specifically, the integrated

circuits used were 3 OP747 for the operational

amplifiers and 3 AD620 for the instrumentation

amplifiers. The signal conditioning circuit weights

less than 15 grams. Its main electrical characteristics

were experimentally checked and are shown in the

next section.

2.2 Signal Recordings

Five recording sessions, of about three hours, were

carried out in healthy volunteers in fast state (>8h).

Subjects were in a supine position inside a Faraday

cage. Firstly the abdominal body surface was

exfoliated to remove dead skin cells to reduce

contact impedance. The abdominal surface was also

shaved in male subjects.

Figure 2, shows the location of electrodes for the

EEnG recordings. The developed flexible electrode

array was placed horizontally 2.5 cm below the

umbilicus, providing three laplacian signals. Three

monopolar Ag-AgCl floating electrodes of 8mm of

sensing diameter were placed 2.5cm above the

umbilicus. Interelectrode distance was also 2.5cm.

Two bipolar recordings of EEnG were obtained from

adjacent monopolar electrodes.

The main sources of physiological interferences

usually present on surface EEnG were

simultaneously recorded. Specifically, ECG was

monitored by Lead 1 using disposable electrodes;

respiration was recorded by an airflow transducer

(1401G Grass), and movements were measured by

means of 3-axis accelerometer (ADXL 335).

Figure 2: Location of electrodes for EEnG recordings.

All signals, except from acceleration signals,

were amplified and band-pass filtered (0.1-100Hz)

by means of commercial bioamplifiers (Grass P511).

A disposable electrode placed on the left ankle of the

FlexibleArrayofActiveConcentricRingElectrodesforSurfaceBioelectricalRecordings-ApplicationtoNon-invasive

RecordingsofEEnG

subject was used as reference for the bioelectrical

recordings. Signals were simultaneously recorded at

a sampling rate of 1kHz.

2.3 Signal Analysis

In order to study the activity of the low-frequency

component of the EEnG i.e. the slow wave, EEnG

and respiratory signals were low-pass filtered

(fc=0.5Hz) and resampled at 4Hz.

The power spectral density (PSD) of signals was

estimated by means of autoregressive parametrical

techniques (AR, order 120). PSD was estimated for

moving windows of 120s every 15s of the recorded

signals. The dominant frequency (DF) over 8 cpm of

the PSD of every window was determined. The

parameter %Resp was defined as the ratio between

the number of windows in which the DF of the

surface signal (bipolar or laplacian) is within the DF

of respiration ±0.5 cpm and the total number of

windows. Similarly, %TFSW is defined as the ratio

of analysed windows whose DF is inside the typical

frequency range of intestinal slow wave (8-12 cpm).

The rest of cases are included in the parameter

%Other.

3 RESULTS

3.1 Active Electrode Array

Figure 3 shows the sensing part of the array of active

concentric ring TCB electrodes. It can be

appreciated that the substrate is flexible enough to fit

the body surface curvature. Moreover the adhesion

of the conductor paste to the substrate was checked

by means of a sticky tape (8915 Filament APT, 3M).

The paste took off after more than 30 cycles proving

the good adherence.

Both sides of the signal preconditioning circuit

of the active electrode array can be seen in figure 4.

The small size and weight of the circuit and the

flexible nature of the array makes it possible to place

this part above the electrodes. With the proper fixing

strategy, the active electrode array could be used for

ambulatory monitoring. Table 1 summarizes the

main electrical characteristics of the developed

signal conditioning circuit. It can be observed that

the battery life is adequate for the recording

sessions, and the CMRR and output noise are also

appropriate for bioelectrical applications.

Figure 3: Implemented flexible array of three TCB

concentric ring electrodes.

Figure 4: Signal preconditioning circuit: bottom side (left)

and top side (right).

Table 1: Main electrical parameters of the signal-

preconditioning circuit.

Cut-off frequency of high pass filter 0.049 Hz

Differential gain at medium freq. 3386 V/V

CMRR at medium frequencies 116 dB

CMRR at 50 Hz 103 dB

Output noise 0.195 mVrms

Battery life 280 min.

3.2 EEnG Monitoring

Figure 5 shows an example of the biosignals

simultaneously recorded. It can be appreciated that,

as expected, the amplitude of the signals picked up

by the ringed-electrodes of the array is smaller than

that of the bipolar recordings with disc electrodes.

Nevertheless, the conventional bipolar recordings

present stronger ECG interference as it can be easily

observed in this figure. In the signals from

concentric ring electrodes the electrocardiographic

interference is almost null. It can only be hardly

appreciated in the signal corresponding to electrode

1 (Lp1) which is placed on the left side of the

BIODEVICES2013-InternationalConferenceonBiomedicalElectronicsandDevices

subject. Regarding the intestinal SW activity,

approximately five waves can be identified on the

laplacian recordings. In bipolar recordings this is

difficult to identify by visual inspection since it

seems they are strongly corrupted by the respiration.

This can also be observed in the example of PSD of

signals shown in figure 6. In bipolar recordings the

dominant frequency (DF) corresponds to the

respiratory frequency; whereas the DF of laplacian

recordings is around 10.5 cpm which fits the normal

SW frequency in human jejunum.

Figure 5: Simultaneous recording of biosignals: Lp1-3:

laplacian signals from the electrode array; Bp1-2: bipolar

signals from disc electrodes; Resp: respiration; ECG:

electrocardiogram

Figure 6: Example of power spectrum density (AR120) of

a 120s window of recorded biosignals.

The results of the %TFSW, %Resp and %Other

parameters, which are presented in table 2, confirm

this behaviour. It is shown that in around 25% of the

signal windows studied, the respiratory interference

masks the intestinal SW activity. In contrast, this

ratio is only around 10% for the laplacian

recordings. Furthermore, around 75% of the cases

the DF of laplacians recordings the DF is in the

frequency range associated to the intestinal SW,

whereas it is about 65% for the conventional bipolar

recordings. Finally to say that the number of cases

whose DF is associated to SW armonics or other

components is slightly higher in laplacian recordings

than in conventional bipolar recordings.

Table 2: Percentage of dominant frequency in the

bandwidths of the different components (mean±standard

deviation); N=3385, Lp1-3: laplacian signals from the

electrode array, Bp1-2: bipolar signals from disc

electrodes.

Channel %TFSW %Resp %Other

Lp1 76,6±8,6 11,3±5,6 12,0±5,8

Lp2 72,8±8,2 9,5±7,0 17,7±6,2

Lp3 71,2±6,9 11,6±8,0 17,2±3,8

Bp1 63,0±12,1 27,3±12,0 9,7±4,8

Bp2 65,6±13,8 25,2±13,8 9,2±6,0

4 DISCUSSION

To the authors’ knowledge, the flexible array of

active concentric ring electrodes presented in this

paper is the first one of these characteristics. Other

authors have developed active concentric ring

electrodes but on rigid substrates (Li et al., 2005);

(Prats-Boluda et al., 2011); (Koka and Besio, 2007).

This new sensor is more comfortable for the subject

under study and provides a better contact since it

adapts to the body surface curvature. Our group has

recently developed other flexible concentric ring

electrodes (Prats-Boluda et al., 2012). However such

electrodes require the screen-printing of three layers,

alternating conductor and dielectric pastes. In

flexible substrates it is very complicated to use bias

between layers, and the solution proposed in the

present work favours an easier manufacturing.

Moreover, in contrast to individual electrodes (Prats-

Boluda et al., 2012); (Li et al., 2005); (Koka and

Besio, 2007), the electrode array developed in this

work is a more compact solution that reduces the

signal preconditioning cost and space, and it is more

suitable for bioelectrical mapping of the body

surface. Furthermore, the modularity of the

developed sensor permits to reuse the signal

conditioning circuit while the sensing part can be

disposed for hygienic reasons.

Signal recording experiences of this work show

that active concentric electrodes of the flexible array

-50

Lp1

(



-50

Lp2

(



-50

Lp3

(



-100

100

Bp1

(



-100

100

Bp2

(



-50

Resp

(



0 5 10 15 20 25 30

-1

Time (s )

ECG

(mV)

PSD (a.u.)

0 4 8 12 16 20 24 28

PSD (a.u.)

Frequency (cpm)

0 4 8 12 16 20 24 28

PSD (a.u.)

Frequency (cpm)

Lp1

Bp1

Lp3

Resp

Bp2Lp2

FlexibleArrayofActiveConcentricRingElectrodesforSurfaceBioelectricalRecordings-ApplicationtoNon-invasive

RecordingsofEEnG

enhance the quality of non-invasive EEnG signals in

terms of electrocardiographic and respiratory

interferences, in comparison to bipolar recordings

with conventional disc electrodes. This is in

agreement with previous studies that used this kind

of electrodes implemented on rigid substrates (Prats-

Boluda, 2011). On one hand, the reduction of

respiratory interference permits to identify more

easily the activity of intestinal slow wave. On the

other hand, the reduction of ECG interference could

help the identification of spike bursts activity. This

could provide more robust systems to non-invasively

monitor intestinal myoelectrical activity which could

bring close the clinical application of this technique.

Nevertheless, this should be confirmed in future

studies.

Moreover, although it has not been tested in this

work, according to other authors (Besio and Chen,

2007); (Soundararajan and Besio, 2005) the

laplacian potential mapping can enhance spatial

sensibility for surface bioelectrical activity. This can

be of great importance for the studies of propagation

maps of cardiac, electroencephalographic or uterine

activity which can provide electrophysiological

information of clinical relevance. The developed

flexible array of active concentric ring electrodes

would be very suitable for these applications.

5 CONCLUSIONS

The flexible array of active concentric ring

electrodes developed in this paper joins the benefits

of laplacian techniques in terms of enhancing spatial

resolution, with the comfort and adaptation to body

surface curvature of conventional disposable

electrodes.

The non-invasive recordings of intestinal

myolectrical signals with this new kind of electrodes

provide enhanced bioelectric signals in terms of

robustness to physiological interferences such as

ECG and respiration, and permit to identify more

easily the intestinal slow wave activity.

ACKNOWLEDGEMENTS

Research supported in part by the Ministerio de

Ciencia y Tecnología de España (TEC 2010-16945)

and by Universitat Politènica de València (PAID

2009/10-2298).

REFERENCES

Besio, W., Aakula, R., Koka, K., and Dai, W., 2006.

Development of a tri-polar concentric ring electrode

for acquiring accurate Laplacian body surface

potentials. Ann.Biomed.Eng 34 426-435

Besio W., and Chen T., 2007. Tripolar Laplacian

electrocardiogram and moment of activation

isochronal mapping. Physiol, Meas. 28 515-529.

Bradshaw, L. A., Richards, W. O., and Wikswo, J. P., Jr.,

2001. Volume conductor effects on the spatial

resolution of magnetic fields and electric potentials

from gastrointestinal electrical activity. Med. Biol.

Eng. Comput. 39 35-43

Chang S. J., Cho E. T., Heo G. S., Park, C. G., Kim M.

W., Chang I. Y., Shin M. K., Cha, K. H., Yeum, C. H.,

Jun, J. Y., 2006. Fasting and postprandial small

intestinal slow waves non-invasively measured in

subjects with total gastrectomy. Gastroenterology

247-252.

Chen, J. D., Schirmer, B. D. and Mccallum, R. W., 1993.

Measurement of Electrical-Activity of the Human

Small-Intestine Using Surface Electrodes. IEEE Trans.

Biomed. Eng. 40 6 598-602.

Fleckenstein P. and Oigaard A., 1978. Electrical spike

activity in the human small intestine. A multiple

electrode study of fasting diurnal variations.

Am.J.Dig.Dis. 23: 776-780.

Garcia-Casado J., Martinez-de-Juan J. L., Ponce J. L.,

2006. Adaptive filtering of ECG interference on

surface EEnGs based on signal averaging. Physiol

Meas. 27: 509-527.

Hjorth, B., 1975. An on-line transformation of EEG scalp

potentials into orthogonal source derivations

Electroencephalogr. Clin. Neurophysiol. 39 526-530

Koka, K. and Besio, W. G., 2007. Improvement of spatial

selectivity and decrease of mutual information of tri-

polar concentric ring electrodes. J.Neurosci. Methods

165 216-222.

Li, G., Wang, Y., Jiang, W., Wang, L. L., Lu, C-Y S., and

Besio, W. G., 2005. Active Laplacian Electrode for the

Data-acquisition System of EHG. Journal of Physics:

Conference Series, 13, 330-335.

Lu, C. C. and Tarjan, P. P. 1999. An ultra-high common-

mode rejection ratio (CMRR) AC instrumentation

amplifier for laplacian electrocardiographic

measurement. Biomed. Instrum. Technol. 33 76-83.

Perrin, F., Pernier, J., Bertrand, O., Giard, M. H., and

Echallier, J. F. 1987. Mapping of scalp potentials by

surface spline interpolation. Electroencephalogr. Clin.

Neurophysiol. 66 75-81.

Prats-Boluda, G., Garcia-Casado, J., Martinez-de-Juan, J.

L., and Ponce J. L. 2007. Identification of the slow

wave component of the electroenterogram from

Laplacian abdominal surface recordings in humans.

Physol. Meas 28 1115-1133.

Prats-Boluda, G., Garcia-Casado, J., Martinez-de-Juan, J.

L., and Ye-Lin, Y. 2011. Active concentric ring

electrode for non-invasive detection of intestinal

myoelectric signals. Med.Eng Phys. 33 446-455.

BIODEVICES2013-InternationalConferenceonBiomedicalElectronicsandDevices

Prats-Boluda, G., Ye-Lin, Y., Ibañez, J., Garcia-Breijo, E.,

Garcia-Casado, J., 2012. Active flexible concentric

ring electrode for non invasive surface bioelectrical

recordings. Meas. Sci. Technol, (to be published).

Quigley, E. M. 1996. Gastric and small intestinal motility

in health and disease. Gastroenterol. Clin.North Am.,

25 113–145.

Soundararajan, V. and Besio, W. 2005. Simulated

Comparison of Disc and Concentric Electrode Maps

During Atrial Arrhythmias. International Journal of

Bioelectromagnetism 7 217-220.

Weisbrodt, N. W. 1987. Motility of the Small Intestine. in

Physiology of the Gastrointestinal Tract, 2nd ed. vol.

1, L. R. Johnson, Ed. New York: Raven Press, , pp.

631–663.

Wu, D., Tsai, H. C., and He, B. 1999. On the estimation of

the Laplacian electrocardiogram during ventricular

activation. Ann.Biomed.Eng 27 731-745.

FlexibleArrayofActiveConcentricRingElectrodesforSurfaceBioelectricalRecordings-ApplicationtoNon-invasive

RecordingsofEEnG