Every time I show someone how to use our Biosemi system they want to know how to measure impedance and seem very worried that they can't. This is especially true of people who have experience with EEG. I'm always a bit unsure about how to answer their questions. Below I've written my latest attempt at explaining why with passive electrodes checking impedance is especially important but with Biosemi active electrodes it is less of an issue (at least that is what I'm assuming). Can you please give me some feedback and more information on this issue? This will be very useful for anyone using the system who may face the impedance issue in dealing with colleagues or reviewers of manuscripts. I apologize if the reasons for this are clearly laid out elsewhere in the technical publications. Unfortunately, I don't have the expertise to really understand those (I have tried) so I'm looking for something on a level (basic college electricity and magnetism. I was a biology major!) that I and other users of the system can grasp more intuitively (but still accurately!).
(1) PASSIVE ELECTRODES - Current flows into/out of the head across the skin/gel/electrode interface and through the electrode wire to/from the amplifier. At various points along this path the current will face some resistance (something that restricts the flow of current). For alternating current (as we have in EEG signals) this resistance is called impedance because in addition to resistance, capaciatance and inductance also affect the ability of AC current to flow. Impedance/resistance is related to voltage and current by Ohm's law; Voltage = Current x Resistance. If the impedance across the skin/gel/electrode interface is high, then this will mean that it will be harder for current to flow to/from the amplifier (especially along a long wire that has lots of chances for interference) and a higher voltage will be necessary in order to get a signal through the wire to the amplifier. Thus, higher impedances result in poorer signal quality because amplitudes will be attenuated. Impedance across the skin/gel/electrode interface is usually lowered by removing the layer of dead skin cells on the surface of the skin.
(2) BIOSEMI ACTIVE ELECTRODES - A "pre-amplifier" is inserted just after the electrode and before the long wire that goes to the amplifier. This amplifier does not give the signal any gain. I'm not sure exactly what happens here? Is common mode (CMS) rejection done here and the difference then transmitted down the wire? Why does the impedance across the gel/skin/electrode interface no longer matter? Why doesn't the same situation as for passive electrodes come up (higher impedance > lower current down the wire > lower measured voltage?) or have I gotten that wrong?
Impedance seems to have a big effect on signal quality in passive systems and thus it needs to be equivalent across channels. Is there some other factor that has a major impact on signal quality in the Biosemi system and does it need to be equivalent across channels. I have noticed that some channels do need to have more gel or a little scratch or two to make the signal APPEAR QUALITATIVELY better (i.e. less noisy, more stable). Is this changing the impedance? It would be nice to have some more QUANTITATIVE measure of what I have done.
I very much appreciate any help that you can give on these issues. I want to understand better how the system is working...
understanding active electrodes and impedance
-
Coen
- Site Admin
- Posts: 1148
- Joined: Fri Mar 26, 2004 7:00 pm
- Location: Amsterdam, Netherlands
- Contact:
EEG currents do not (significantly) flow via the electrodes because the input impedance of all current biopotential amplifiers is very high. So, typical electrode impedances (smaller than a few hundred kOhm) do not influence measured EEG voltages.
Interference currents flow via the electrode impedances due to stray capacitances between interference sources (e.g. mains wiring), and the electrode wires and the subject body, see http://www.biosemi.com/publications/pdf ... uction.pdf . This current flow causes interference voltages proportionally to the electrode impedances. With passive electrodes, it is necessary to lower the electrode impedances to the kOhm range, to archive an acceptable level of interference in typical measurement situations.
Active electrodes provide impedance transformation on the electrode: the input impedance is very high (so the EEG voltages are not influenced, even with high electrode impedances), while the output impedance is very low (< 1 Ohm). Consequently, the interference currents now flow via very low impedances (the output of the active electrode), and cannot generate significant interference voltages anymore. In addition, the electrode impedance does not longer affect the level of interference.
Because the actual electrode impedance is not a very important variable, when active electrodes are used, the level of DC offset is used as an alternative indicator for the quality of the electrode contact.
Best regards, Coen (BioSemi)
Interference currents flow via the electrode impedances due to stray capacitances between interference sources (e.g. mains wiring), and the electrode wires and the subject body, see http://www.biosemi.com/publications/pdf ... uction.pdf . This current flow causes interference voltages proportionally to the electrode impedances. With passive electrodes, it is necessary to lower the electrode impedances to the kOhm range, to archive an acceptable level of interference in typical measurement situations.
Active electrodes provide impedance transformation on the electrode: the input impedance is very high (so the EEG voltages are not influenced, even with high electrode impedances), while the output impedance is very low (< 1 Ohm). Consequently, the interference currents now flow via very low impedances (the output of the active electrode), and cannot generate significant interference voltages anymore. In addition, the electrode impedance does not longer affect the level of interference.
Because the actual electrode impedance is not a very important variable, when active electrodes are used, the level of DC offset is used as an alternative indicator for the quality of the electrode contact.
Best regards, Coen (BioSemi)
Re: understanding active electrodes and impedance
Hello,
In reference to BioSemi's active electrodes, the original poster asks the question "Why does the impedance across the gel/skin/electrode interface no longer matter?"
I've been reading through the forums and notice that there are many descriptions about why 'electrode impedance' is negligible in the active two system. However, I'm still struggling to understand how differences in impedances at each channel that would specifically be caused by variations in gel/skin properties (prior to reaching the electrode) are taken into account in the recording of EEG data. Or, if this is not an important variable to consider, why not?
My apologies if my limited understanding of what has already been said makes this a redundant question. I do really appreciate any light that can be shed on this.
Best,
Matthew
In reference to BioSemi's active electrodes, the original poster asks the question "Why does the impedance across the gel/skin/electrode interface no longer matter?"
I've been reading through the forums and notice that there are many descriptions about why 'electrode impedance' is negligible in the active two system. However, I'm still struggling to understand how differences in impedances at each channel that would specifically be caused by variations in gel/skin properties (prior to reaching the electrode) are taken into account in the recording of EEG data. Or, if this is not an important variable to consider, why not?
My apologies if my limited understanding of what has already been said makes this a redundant question. I do really appreciate any light that can be shed on this.
Best,
Matthew
-
Coen
- Site Admin
- Posts: 1148
- Joined: Fri Mar 26, 2004 7:00 pm
- Location: Amsterdam, Netherlands
- Contact:
Re: understanding active electrodes and impedance
EEG data consist of voltages (potential differences between the electrode sites). By Ohms law, the electrode impedances can only cause error voltages when electrical current flows through the electrodes. Currents may flow through the electrodes as a result of capacitive coupling of the electrode wires with mains supply wires, due to input bias currents of the amplifier, or due to a low input impedance of the amplifier.
The active electrode setup prevents significant currents to flow through the skin-gel-electrode interface because it eliminates interference currents to flow through the electrode (these currents flow into the low-impedance output of the active electrode buffer amplifier), because the input bias currents are minimal (CMOS input), and because the input impedance is maximal (minimal wire length between electrode and first amplifier stage offers lowest possible input capacitance).
Best regards, Coen (BioSemi)
The active electrode setup prevents significant currents to flow through the skin-gel-electrode interface because it eliminates interference currents to flow through the electrode (these currents flow into the low-impedance output of the active electrode buffer amplifier), because the input bias currents are minimal (CMOS input), and because the input impedance is maximal (minimal wire length between electrode and first amplifier stage offers lowest possible input capacitance).
Best regards, Coen (BioSemi)
Re: understanding active electrodes and impedance
My take on this
Active electrodes:
Amplifier close to the signal origin skin-side, so the connecting cable to the unit will transport a signal already amplified, which will be less degraded by the existing magnetic\electric coupling to the cable. Moreover since the output resistance of the active electrodes´ amplifiers is so low this will reduce the coupling effect of the magnetic\electric noise on the cables. This simply reflects the fact that in a acquisition chain the noise factor of the front-end components contribute more to the overall noise factor, than the remaining component´s noise factor.
Passive electrodes.
Amplifiers located inside the acquisition units closer to the units background electrical/magnetic noise and further away from the signal origin (skin) so the low level biosignal low amplitude will "compete" with the coupling magnetic/electrical interference on the electrode cables which "see" the high-input impedance of the amplifiers. Driven Shield Inputs and guard techniques reduce this interference but I believe very few experimental units do this.
Passive Electrode amplifiers with input impedance as high as the active electrode amplifiers.
Electrode impedance estimation in passive electrodes:
Easily achieved with passive electrodes by reverting the signal directionality in the electrode cables: instead of receiving the biosignal, the unit sends to the body a sinusoidal current around 1 microAmp (not DC to avoid polarization) and measures the developed voltage. Impedance is estimated by the Omhs law: R=V/I. Return current is done via reference or ground. Normally this procedure is made prior to acquisition and switched off thereafter, as it would interfere with acquisition reducing the CMRR.
Electrode impedance estimation in active electrodes:
The Active two Biosemi system does not have a patient ground per se, so looks difficult to be able to have the return current for the mentioned above 1 microAmp current: Moreover, the active electrode amplifiers in the skin side, would have to be (for impedance measurement) reconfigured to work on "reverse" and inject a current in the patient, or, at least, let this current pass to the patient, which I believe is not possible here. I believe there is no signal control communication protocol between the Biosemi unit and their active electrodes, which just "send" the amplified bio-signal to the unit, so can not play any role besides that of an amplifier?
The conjunction of these two factors seems to make the R=v/I impedance measurement in the Biosemi system a tricky thing, if not impossible, thus the reliance in the dc offset levels via CMS to estimate good skin contact.
As a Biosemi user this is my understanding of the impedance issue...
Active Electrodes:
The main drawback it is (1)they are not disposable which puts hygiene problems, still unsolved with some conflicting remediation methods , (2) intrinsically, as far as interface electrode-skin contact is concerned their noise may be higher than a disposable low noise one. Ageing and poor maintenance of the active electrodes induces pop noise, so basically we are trading low noise associated with active systems with electrode noise (3) relatively expensive to replace (4) Relatively more fragile then passive electrodes
Active electrodes:
Amplifier close to the signal origin skin-side, so the connecting cable to the unit will transport a signal already amplified, which will be less degraded by the existing magnetic\electric coupling to the cable. Moreover since the output resistance of the active electrodes´ amplifiers is so low this will reduce the coupling effect of the magnetic\electric noise on the cables. This simply reflects the fact that in a acquisition chain the noise factor of the front-end components contribute more to the overall noise factor, than the remaining component´s noise factor.
Passive electrodes.
Amplifiers located inside the acquisition units closer to the units background electrical/magnetic noise and further away from the signal origin (skin) so the low level biosignal low amplitude will "compete" with the coupling magnetic/electrical interference on the electrode cables which "see" the high-input impedance of the amplifiers. Driven Shield Inputs and guard techniques reduce this interference but I believe very few experimental units do this.
Passive Electrode amplifiers with input impedance as high as the active electrode amplifiers.
Electrode impedance estimation in passive electrodes:
Easily achieved with passive electrodes by reverting the signal directionality in the electrode cables: instead of receiving the biosignal, the unit sends to the body a sinusoidal current around 1 microAmp (not DC to avoid polarization) and measures the developed voltage. Impedance is estimated by the Omhs law: R=V/I. Return current is done via reference or ground. Normally this procedure is made prior to acquisition and switched off thereafter, as it would interfere with acquisition reducing the CMRR.
Electrode impedance estimation in active electrodes:
The Active two Biosemi system does not have a patient ground per se, so looks difficult to be able to have the return current for the mentioned above 1 microAmp current: Moreover, the active electrode amplifiers in the skin side, would have to be (for impedance measurement) reconfigured to work on "reverse" and inject a current in the patient, or, at least, let this current pass to the patient, which I believe is not possible here. I believe there is no signal control communication protocol between the Biosemi unit and their active electrodes, which just "send" the amplified bio-signal to the unit, so can not play any role besides that of an amplifier?
The conjunction of these two factors seems to make the R=v/I impedance measurement in the Biosemi system a tricky thing, if not impossible, thus the reliance in the dc offset levels via CMS to estimate good skin contact.
As a Biosemi user this is my understanding of the impedance issue...
Active Electrodes:
The main drawback it is (1)they are not disposable which puts hygiene problems, still unsolved with some conflicting remediation methods , (2) intrinsically, as far as interface electrode-skin contact is concerned their noise may be higher than a disposable low noise one. Ageing and poor maintenance of the active electrodes induces pop noise, so basically we are trading low noise associated with active systems with electrode noise (3) relatively expensive to replace (4) Relatively more fragile then passive electrodes