=Paper=
{{Paper
|id=Vol-2473/paper36
|storemode=property
|title=Comparison of Dry Electrodes for Mobile EEG System
|pdfUrl=https://ceur-ws.org/Vol-2473/paper36.pdf
|volume=Vol-2473
|authors=Marianna Koctúrová,Jozef Juhár
|dblpUrl=https://dblp.org/rec/conf/itat/KocturovaJ19
}}
==Comparison of Dry Electrodes for Mobile EEG System==
https://ceur-ws.org/Vol-2473/paper36.pdf
Comparison of dry electrodes for mobile EEG system
Marianna Koctúrová, Jozef Juhár
Dept. of Electronics and Multimedia Communications,
FEI, Technical University of Košice
Košice, Slovak Republic
marianna.kocturova@tuke.sk,
jozef.juhar@tuke.sk,
Abstract: The main objective of this study was to eval- or assisting people. BCI-based assistance may include ap-
uate two types of dry EEG electrode. In the paper, we plications to improve the nervous system or restore nerve
describe the comparison of two comb electrodes. The first bonds in the case of paralysis [2].
was an electrode based Ag-AgCl alloy and the second was The EEG device, which would allow ordinary people to
electrode based on a flexible conductive polymer. Testing use the BCI interface in everyday life, has several condi-
of these electrodes was performed based on the need to tions. These conditions are suitable device design, ease
increase convenience when measuring EEG signals while of use in non-clinical settings, comfort, painlessness, and
maintaining the same signal characteristics. cleanliness. Medical EEGs use conductive gel electrodes
and are made in the form of an elastic cap. The mo-
bile EEG should be usable without the need for shaving
1 Introduction the head and comfortably enough, so dry EEG electrodes
should be used. For these reasons, we have performed the
Most brain-computer interfaces (BCI) are based on brain experiment of the using and properties of dry electrodes.
wave recording using electroencephalography (EEG).
EEG technology is a non-invasive method for recording
signals derived from brain activity. EEG uses electrodes 2 Materials and methods
deposited on the human scalp to capture the signal that
passes through the skull. This signal is considerably weak- In the experiment, two types of electrodes were compared.
ened compared to the original, so the EEG device must be As the first step, the resistance of the electrodes was eval-
suitably designed to capture it [1]. uated using the Volt-Ampere method. In the second part,
In general, there are several types of EEG electrodes. the quality of the signals measured with these electrodes
Disc electrodes used in medicine require the use of con- was compared.
ductive electrode for optimal impedance and data quality.
However, the use of a gel has serious disadvantages and
2.1 EEG headset
problems which are particularly noticeable when captur-
ing EEG signals in real-life conditions by using EEG de- The OpenBCI headset was used to measure the EEG sig-
vice by laymen. Therefore, so-called dry electrodes that nal. The headset is designed as a plastic 3D printed con-
do not require the presence of any additive are used for struction, that allows place electrodes up to 35 different
BCI applications. The use of dry electrodes that do not re- positions by standard 10/20 configuration system. In the
quire gel is often very advantageous as it provides a quick experiment, the brain signal from 10 locations of frontal
setting of the device without time-consuming preparation, and temporal lobes was measured.
but often brings new problems such as comfort and signal The headset works wirelessly. The measured data is sent
quality. Recent studies focus on the use of EEG signals via the Bluetooth 4.0 wireless communication protocol to
in mobile BCI applications. Such applications should be the USB dongle receiver. The headset can also store data
based on the use of such EEG devices that should be as directly on a microSD card when the device is not con-
comfortable as possible and should be easy to use for the nected to any wireless receiver. The entire system is pow-
individual. ered by batteries, providing greater patient electrical safety
Dry electrodes have the ability to make EEG technol- and portability [3].
ogy available for mobile applications. Mobile EEG ap- The input impedance of the amplifiers in the headset is
plications and EEG devices can make life easier or more 500MΩ. The lead resistance of the electrode is negligible
comfortable for many people. There are several areas of compared to the large impedance at the amplifier input [4].
use of Brain to computer interfaces (BCI) such as games
2.2 Post dry electrode
Copyright c 2019 for this paper by its authors. Use permitted un-
der Creative Commons License Attribution 4.0 International (CC BY The first dry EEG electrode TDE-200 was used. The elec-
4.0). trode is also known as Post electrode, shown in Figure 1.
Figure 1: TDE-200 dry EEG-electrode (Post electrode)
Figure 3: Electrode measurement
3 Electrode electrical properties
measurement
Figure 2: Datwayer Brush electrode Due to the low resistance and voltage values of the elec-
trodes, the Volt-Ampere method to measure the resistance
was used to ensure the most accurate result. The measure-
It is a dry electrode, made of Silver-silver chloride (Ag- ment using a copper plate was performed, as it is shown in
AgCl) alloy with a diameter of 10 mm. This electrode Figure 3.
is specific in that it contains 12 pins to improve contact The amount of voltage and the current passing through
through fur or hair. The pins are 2 mm long to provide the electrodes was measured. Because of the small size
good contact with the skin surface through the hair. They and irregular shape of the electrode electrodes, we used the
provide accurate and clear transmission of surface biopo- copper plate on which we placed the electrode to measure.
tentials. The electrodes connection is provided by a screw Then we measured the current passing through the plate
to which the conductive cable is then attached [5]. and the end of the electrode and the voltage between these
The advantage of the electrodes is the low resistance points.
due to the metal composition. The use of electrodes does
not require the addition of a moisturizing gel to improve
conductivity or to remove hair from the measured head 3.1 Electrical features of the Brush electrode
area. Thanks to the small pins, the electrode can reach the
skin surface. As the first was measure the Post electrode. The Post elec-
The disadvantage is the painful setting of the electrodes trode is made of Silver-silver chloride (Ag-AgCl), there-
before the measurement and the occasional subjective pain fore, lower resistance was assumed. This assumption was
even during longer measurements. Pins often push too confirmed by measurement.
much on the skin, but leave no injuries. The measurements showed electrode voltage values of
2.6 to 4.3mV and current values were measured in the
range of 101-102mA. The electrode resistance was then
2.3 Brush electrodes calculated by the Volt-Ampere method. The resistance val-
ues were from 26 to 43mΩ.
The second tested sample was dry Brush electrode by
Datwyler, shown in Figure 2. The electrode is designed
for better comfort. The electrodes are based on a flexible 3.2 Electrical features of the Post electrode
conductive, elastic main body with a conductive coating
covering the contact area, ensuring comfort during moni- Brush electrode measurement was performed to compare
toring and setting of the headset. Brush electrode also has the values of resistance. In the experiment, the voltage
small pins, but these are soft and movable due to that they between the ends of the electrode was 2,6V . The Brush
are made of conductive polymer. There are 15 pins with a electrode current values were in the range of 70-78mA,
length of 5 mm and the contact area has a diameter of 12 which depended on the Brush electrode being pressed to
mm. They are attached to standard snap lead cabling. the copper plate. The average current value was 74mA.
The advantage of the Brush electrodes is that they are Resistance was calculated from the mean values by the
designed for dry signal acquisition, so they do not need volt-ampere method. Its values were calculated as 35Ω.
moisturizing gels or hair removal for using. Thanks to Since the electrode is on the entire surface of the poly-
longer pins they can work through the hair. Elastic ma- mer, its resistance values are relatively high. Approximate
terial ensures an easier setting and painless measurement resistance values could also be measured directly by the
on the skin [6]. multimeter.
Figure 5: Skin irritation for short term use of dry elec-
trodes. In the top: Post electrode with pins of 2 mm length;
Figure 4: Electrode configuration
At the bottom: Brush electrode with pins of 5 mm length;
4 EEG data acquisition
bandwidth was filtered out of the signal. The alpha wave
frequency is 8-12Hz. Alpha waves are created in a state of
The next step was to measure the EEG signals. An Open-
wakefulness when a person feels relaxed. By filtering this
BCI headset has been used to record signals. Signals were
bandwidth some artefacts have been removed and only the
measured to compare the values and properties of signals
clear brain signal from the subject’s relaxation state has
obtained from two different electrodes.
remained [8].
Post and Brush electrodes were placed in the EEG head-
Individual hemispheres and brain parts are responsible
set at the same time to observe signal similarities under the
for the various processes. Nevertheless, we can compare
same conditions. Two session signals were recorded dur-
the values of signals in opposite positions. Artefacts or
ing the experiment.
some brain activity with stronger pulses can also be oc-
Based on the 10-20 configuration system, in the first
curred on opposite sides of the hemisphere simultaneously.
session Brush electrodes were placed on the right hemi-
In the Matlab programming environment, signals obtained
sphere of the head and Post electrodes placed on the left
from individual electrodes were compared. In the first
hemisphere. In the second session, the electrode positions
step, the voltage levels on the both electrodes were com-
were reversed. Figure 4 shows the electrode positions of
pared. The comparison was always performed between
the experiment in 10-20 configuration system [7]. The po-
the electrodes at the same positions on the opposite hemi-
sitions on the left hemisphere are marked in blue and the
spheres.
yellow ones indicate the right. 10 electrodes were used
for measurement, 5 Post and 5 Brush electrodes for each The best comparison can be made on the forehead on the
session. electrodes Fp1 and Fp2. This part transmits an electrical
The brain waves of a subject in the state of relaxation signal coming from visually evoked potentials and from
were recorded. The subject sat still in a comfortable po- eye movement. Therefore, comparing these signals is ap-
sition while playing the video. The subject watched the propriate for the case in our experiment where the subject
video was with a light concentration. Each session lasted watched the video while recording the EEG.
60 seconds. In the state of relaxation and light concentra- The correlation of the raw signals from the first session
tion, in the brain are created so-called alpha waves [8]. was 0.92 on the Fp1 and Fp2 electrodes and after filtering
Figure 5 shows skin irritation. It can be seen that the the alpha bandpass, the correlation was 0.87. In the second
Brush electrode leaves the skin reaction longer than the session, the correlation between Fp1 and Fp2 signals was
Post electrode. The pictures were captured after 10 min- 0.98 and 0.97 on the alpha frequency band.
utes of EEG recording. After 30 minutes almost no re- Figures 6 and 7, show the simultaneous signal plot of
action was seen in the Post electrode application area. In the Alpha wave EEG signal at electrodes Fp1 and Fp2 dur-
contrast, the Brush electrode left a visible sign after 30 ing the first and second session. From the EEG measure-
minutes. ments, we can observe the signal sequence similarity on
the electrodes Fp1 and Fp2.
In Figures 8 and 9, show the comparison of the fre-
5 Signal comparison quency spectres. The Brush electrode signal is displayed
in blue and the Post electrode signal is displayed in red.
In the Matlab programming environment, the signals were The Figures 8 represents frequency spectre of signal from
modified from the default format in which they are saved, the first session, where at position Fp1 there was the Brush
to matrix for ease of evaluation. Then the alpha frequency electrode and at Fp2 was Post electrode. The Figure 9 rep-
Figure 9: Session 2 - The frequency spectrum of alpha
Figure 6: 10 seconds of Alpha waves of EEG signal on
wave signal on electrodes Fp1 and Fp2
electrodes Fp1 and Fp2
the electrodes to the OpenBCI hardware it has been shown
that the use of a higher resistance electrode does not affect
the measurement of the brain signal. The use of a suit-
able amplifier in the EEG headset smoothes the difference
between electrode resistances.
Acknowledgment
The research presented in this paper was supported by
the Ministry of Education, Science, Research and Sport
Figure 7: 10 seconds of Alpha waves of EEG signal on of the Slovak Republic under the research project VEGA
electrodes Fp1 and Fp2 1/0511/17, by the Cultural and Educational Grant Agency
of the Slovak Republic under grant No. 009TUKE-4/2019
and by the Slovak Research and Development Agency
project No. APVV-SK-TW-2017-0005.
References
[1] M. Koctúrová and J. Juhár, “Eeg based voice activity detec-
tion,” in 2018 16th International Conference on Emerging
eLearning Technologies and Applications (ICETA), pp. 267–
272, IEEE, 2018.
[2] J. Wolpaw and E. W. Wolpaw, Brain-computer interfaces:
Figure 8: Session 1 - The frequency spectrum of alpha principles and practice. OUP USA, 2012.
wave signal on electrodes Fp1 and Fp2 [3] O.-O. S. B. Tools, “Openbci.com. retrieved 24 february
2018,” 2018.
resent frequency spectre from the second session, where [4] T. Instruments, “Ads1299-x low-noise, 4-, 6-, 8-channel, 24-
bit, analog-to-digital converter for eeg and biopotential mea-
electrode positions were replaced.
surements,” Jul-2012.[Online]. Available: http://www. ti.
The increase in frequency spectrum power on Brush com/lit/ds/symlink/ads1299. pdf.[Accessed: 12-May-2017],
electrodes was on average 20% in the first session. In the 2017.
second, the increase with the use of Brush electrodes was [5] G. Cañadas, C. Dell’Aquila, A. Garces, and E. Laciar, “Vali-
on average 10%. dation of a wireless and portable eeg acquisition system with
dry electrodes,” in World Congress on Medical Physics and
Biomedical Engineering 2018, pp. 833–837, Springer, 2019.
6 Results and Conclusion
[6] “www.datwyler.com,” 2019.
Measurements of the electrodes electrical properties [7] V. Jurcak, D. Tsuzuki, and I. Dan, “10/20, 10/10, and 10/5
showed significantly higher Brush electrode resistance systems revisited: their validity as relative head-surface-
values. From the comparison of EEG signals, it can be based positioning systems,” Neuroimage, vol. 34, no. 4,
pp. 1600–1611, 2007.
stated that although the Brush electrodes is made of a con-
ductive polymer and has a higher resistance, the signal [8] P. Pramanick, Classification of electroencephalogram (EEG)
signal based on fourier transform and neural network. PhD
values obtained by it are comparable to those of the sig-
thesis, 2013.
nal obtained by the metal Post electrode. By connecting