EEG Electrodes and Galvanic Voltages
The combination of electrode material, paste, and substances on the skin surface turn every EEG electrode into a little battery that generates a direct voltage of up to several hundred millivolts.
This is one reason for the fact that EEG amplifiers are always AC-amplifiers - the direct current of these batteries is not amplified (note: nowadays many EEG amplifiers are indeed DC-coupled; the DC component of the signal is then removed by the software).
But why are we interested in the direct voltage anyway?
Direct Voltage - So What?
These direct voltages amount to several hundred millivolts and are therefore up to a factor of ten thousand higher than a typical EEG signal.
If the direct voltage were nice and stable, we would indeed not care about it. In addition to that, we expect all electrode contacts to produce the same direct voltages, so that our differential amplifier would not notice them.
Unfortunately, the real world works differently. Not only do the direct voltages vary between each other, they also vary with time. A typical distribution is shown here. It is not unusual for the differential voltage between two electrodes to be several hundred millivolts shortly after the electrodes have been applied. This voltage typically changes considerably within ten to twenty minutes - also in the range of several hundred millivolts. This change is picked up by an AC-amplifier if it is quick compared to the low cut-off frequency. Example: The EEG NeuroAmp has an effective low cut-off frequency of 0.08Hz (first-order filter). Events with a time constant of over one minute are therefore still significantly amplified.
But that´s not all. The distribution of the galvanic voltages toward a more stable equilibrium is by no means smooth - it is in fact more comparable to a steep and uneven path. These "bumps" significantly influence the EEG recording in a negative way. An irregularity, such as a noise level of only one percent on top of a "direct" voltage of 100 mV, amounts to one millivolt, which is still at least one order of magnitude above the amplitude of a typical EEG signal.
There are two ways of dealing with these facts:
1. We don´t care about the processes ocurring at very low frequencies. The EEG amplifier or the software simply eliminate the low-frequency noise by means of a high-pass filter.
2. We use electrodes with very low galvanic voltages that produce only very little voltage noise.
If we work at frequencies below one Hertz, we need to go for solution 2. We have thoroughly investigated different electrode materials and electrode-paste combinations and have come to the conclusion that sintered silver/silver chloride electrodes are the best. The active part of these electrodes consists completely of silver/silver chloride, not just the surface. Therefore, abrasion of the surface, as always happens during normal use, does not result in deterioration of electrical properties of the electrodes. But there are also drawbacks: These electrodes are fairly costly and the electrode material is mechanically and electrically sensitive. The active material has to be enclosed in a plastic housing and may not come into contact with metal ions or even tap water.
But, in addition to the advantage of the extremely high signal quality, even at low frequencies, these electrodes have the advantage of having a relatively long lifetime if handled correctly.
Of course we offer the electrodes that we liked in our tests and evaluations:
- Set of three or five electrodes (BEE Medic)
- Mushroom electrodes for the SuSaCap
- Single electrodes and earclips (GVB, Germany or Invivo, USA)
Go to the electrode offerings page here.
For more information: