Sara Nazari Asl , Frank Ludwig and Meinhard Schilling

Noise properties of textile, capacitive EEG electrodes

De Gruyter | Published online: September 12, 2015

Abstract

The rigid surface of the conventional PCB-based capacitive electrode produces an undefined distance between the skin and the electrode surface. Therefore, the capacitance introduced by them is uncertain and can vary from electrode to electrode due to their different positions on the scalp. However, textile electrodes which use conductive fabric as electrode surfaces, are bendable over the scalp. Therefore, it provides a certain value of the capacitance which is predictable and calculable accurately if the effective distance to the scalp surface can be determined. In this paper noise characteristics of textile electrodes with different fabric sizes as electrode’s surface and capacity calculations related to each size are presented to determine the effective distances for each electrode size.

1 Introduction

Capacitive electrodes for the evaluation of human body bioelectric potentials are a very attractive alternative to conventional galvanically coupled electrodes for long term diagnostic applications especially when the signals have to be measured through insulating materials like cloth. These electrodes are always constructed in a similar way based on ideas which have been proposed already in the middle of the last century [1, 5]. In the last years many research groups have developed new applications based on these electrodes [2, 3, 6]. We concentrated in our work on the application in a multichannel helmet for electroencephalography (EEG) application e.g. in a brain-computer interface [4]. The common geometric arrangement of the first stage of a capacitive electrode is depicted in Fig. 1.

A capacitive electrode for bioelectric potential measurements consists of a piece of conductive material which forms a capacitor together with some part of the body surface. The electric charges on the body surface caused by the electric body potentials interact with corresponding charges in the conductive electrode which is placed above the body surface separated by some insulating material. As insulator either air or any other dielectric can be employed. Therefore, we will have a capacity Ce of

(1) C e = є 0 є r A e d .

Here є = єrє0 is the permittivity of the dielectric between capacitor’s layers, d is the distance between them, and Ae is the area of the layers.

Figure 1 Electronic network circuit of the electrode determining the frequency response

Figure 1

Electronic network circuit of the electrode determining the frequency response

Together with the input resistance Rin of the impedance converter stage the capacitance is part of an electronic high-pass filter. Normally, the cut-off frequency of the high-pass filter fc which is given by

(2) f c = 1 2 π · R i n · C e

is chosen at about 0.1 Hz in agreement with the bandpass filters used commonly in electrocardiography or electroencephalography.

Here Rin is the bias resistor in parallel with the input resistor of the amplifier which is almost equal to the bias resistor because the input impedance of the amplifier is larger than the bias resistor.

2 Method

We used stretchable conductive fabric which is made from interwoven copper wires. The fabric is not solderable, therefore, we used conductive glue to paste it to the main electronic circuit. For these measurements a PCB plate is used as second side of the input capacitor which is connected to the input and lays underneath the fabric, thus forming the input capacitance Ce.

According to equation (2) to have a lower cut-off frequency, the input resistance Rin and input capacity Ce have to be chosen high enough to let the physiologically interesting frequency components pass. The use of textile electrodes with fabric surface causes smaller average effective distances, thus increases the capacitance compared to stiff PCB electrodes because of better flexible attachment to the body surface.

In this paper we measured the transfer function and noise spectra of electrodes with areas of the fabric electrodes of 64 cm2, 32 cm2, and 16 cm2.

3 Results

In figure 2, the transfer function of the circuit with different areas of the fabric is depicted, the input resistor Rin is 1 in all measurements.

Figure 2 Transfer function of textile electrode with different sizes of the fabric

Figure 2

Transfer function of textile electrode with different sizes of the fabric

From figure 2 we obtain a 3dB cut-off frequency of 0.23 Hz for an area of 64 cm2, 0.52 Hz for an area of 32 cm2, and 0.90 Hz for an area of 16cm2. The corresponding capacitance values for each of the areas can be determined from equation (2). We find Ce(64 cm2) = 0.688 nF, Ce(32 cm2) = 0.305 nF, Ce(16 cm2) = 0.176 nF.

Figure 3 shows the dependence of the cut-off frequency and the electrode capacity on the area of the electrode surface. According to equation (1) and (2), Ce should be proportional to the fabric area while the cut-off frequency should be inversely proportional to the area.

Taking these capacity values and assuming air as dielectric between the surfaces of the capacitors, the distances d were calculated by equation (1) to be about 0.1 mm for all cases. However, looking at figure 3 one sees that the data points for the capacitance vs. area do not lie on a straight line as expected for a constant distance d. To estimate the uncertainty in the distance d, we repeated measurements of the transfer function after assembling and disassembling the capacitors. For the fabric electrode with 16 cm2, we determined the standard deviation of the cutoff frequency as 0.25 Hz. With

(3) δ C = 1 2 π R f ¯ 3 dB 2 δ f 3 dB

one can calculate the uncertaintyC to 0.034 nF with C ¯ = 0 , 177 nF. The uncertainty in the distance between electrodes is determined by the uncertainty in the fabric area δA and the contribution from the uncertainty δC. It is found that the contribution to the uncertainty in distance d of δC by far exceeds that of δA. Thus, the uncertainty in the distance can be calculated by

(4) δ d = є r є 0 A C ¯ 2 δ c

resulting in an uncertainty valued = 0.015 mm.

Figure 3 Cut-off frequency fc and input capacity Ce versus area of the fabric, error bars were determined with equation (3).

Figure 3

Cut-off frequency fc and input capacity Ce versus area of the fabric, error bars were determined with equation (3).

Figure 4 Equivalent noise model circuit

Figure 4

Equivalent noise model circuit

The voltage noise spectra of the electrodes with different areas of fabric are depicted in Fig. 5. The simplified noise model of the electrode is shown in fig. 4. The effect of the noise of the resistor at the output can be calculated by

(5) S v , R t o o t a l = 1 1 + j ω R C 4 k R T δ f

in which k is Boltzmann constant, T is temperature, R is the input resistor, C is input capacitor and δf is the bandwidth. The total noise of the electrode is

(6) S v , T o o t a l = S v , R t o o t a l 2 + S v , o p 2

The white noise of the spectra is the noise of the amplifier Sv, op and slope is caused by the characteristics of the high pass filter. As it is clear in Fig. 5 while the white noise level is the same for all electrode areas, the low-frequency noise decreases with increasing the area. This is expected because cut-off frequency of high-pass filter for larger areas is smaller consequently less noise at lower frequency can pass through it.

4 Conclusion

The cut-off frequency of the high-pass filter of the capacitive electrodes is determined by the input capacity which is received from the electrode surface parallel to the skin surface and the input resistance. In order to have a lower cut-off frequency, the value of input resistor or capacitor should increase. In rigid surface electrodes, an increase in the capacity by increasing the surface of the electrode is not possible because the area of the electrode surface which has a small distance to the scalp is rather small and unpredictable. However, by using conductive fabric as an electrode surface it is possible to increase the area quite considerably while still maintaining the same distance between layers over the whole surface which consequently results in a definite capacity in the input. Therefore, it is possible to shift the cut-off frequency to lower frequencies by increasing the size of fabric in textile electrodes. Simultaneously the voltage noise at lower frequencies is reduced. The distance measured in this work (d = 0.1 mm) is in rough agreement with what we measured for the thickness of the fabric. However, it is clear that more measurement with different fabric should be done to predict what distance should be expected for new unknown fabric.

Figure 5 Voltage noise spectrum of textile electrode with different sizes of the fabric and the simulated noise for 64 cm2

Figure 5

Voltage noise spectrum of textile electrode with different sizes of the fabric and the simulated noise for 64 cm2

Acknowledgment

The author S.N. gratefully acknowledges the financial support by the Braunschweig International Graduate School of Metrology (B-IGSM)(PhD Scholarship,S.N.).

Author’s Statement

    Conflict of interest: Authors state no conflict of interest. Material and Methods: Informed consent: Informed consent has been obtained from all individuals included in this study. Ethical approval: The research related to human use has been complied with all the relevant national regulations, institutional policies and in accordance the tenets of the Helsinki Declaration, and has been approved by the authors’ institutional review board or equivalent committee.

References

[1] Geddes L, Baker L. Principles of applied biomedical instrumentation. Wiley: New York 1989 3rd ed. Search in Google Scholar

[2] Harland C, Clark T, Peters N, Everitt M, Stiffell P. A compact electric potential sensor array for the acquisition and reconstruction of the 7-lead electrocardiogram without electrical charge contact with the skin. Physio. Meas.2005; 26: 939–950. Search in Google Scholar

[3] Lim Y, Kim K, Park K. ECG measurement on a chair without conductive contact,” IEEE Trans. Biomed. Eng. 2006: 53, 956–959. Search in Google Scholar

[4] Oehler M, Neumann P, Becker M, Curio G, Schilling M. Extraction of SSVEP signals of a capacitive EEG helmet for human machine interface. Vancouver: Proc. 30th Annual Conf. IEEE EMBS 2008: 4495–4498. Search in Google Scholar

[5] Richardson P. The insulated electrode: a pasteless electrocardiographic technique. 20th Annual Conference on Engineering in Medicine and Biology 1967; 15.7. Search in Google Scholar

[6] Schommartz A., Eilebrecht B., Wartzek T., Walter M., Leonhardt S. Advances in Modern Capacitive ECG Systems for Continuous Cardiovascular Monitoring. Acta Polytechnica 2011: 51. 5, 100–105. Search in Google Scholar

Published Online: 2015-9-12
Published in Print: 2015-9-1

© 2015 by Walter de Gruyter GmbH, Berlin/Boston

This article is distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.