Page 1 of 1

DRL gel leak

Posted: Tue Mar 07, 2017 3:37 pm
by Giuliano
Hi,

I am using Biosemi Active Two with 64 electrodes to record EEG during tactile stimulation of the index finger though a piezoelectric device. DRL and CMS electrodes are placed in the occipital zone.
During the cap montage (and also during the subsequent analysis of the data) it is clearly visible on the electrode P2 a sinusoidal waveform with the same frequency of the stimulation (22 Hz in our case).
I think this is due to a leakage between the tactile device and the measuring system but it is not clear for me why this signal appear only in one electrode.
Could this be due to a gel bridge between the DRL and P2 (as they are very close each other)?
We repeated the measurement in different day and on different subject so I think it's quite difficult that every time we put too much gel causing a bridge, but I have no other explanation for this weird behaviour.

Best,
Giuliano

Re: DRL gel leak

Posted: Thu Mar 09, 2017 1:28 pm
by Coen
High interference on a single electrode adjacent to the DRL is usually caused by a gel bridge. Check by swapping electrodes between different locations (holders).

Use just enough gel to fill the holder, but not more. Prevent shifting the cap on the head after the gel is applied.

Best regards, Coen (BioSemi)

Re: DRL gel leak

Posted: Fri Mar 10, 2017 11:33 am
by Giuliano
The problem is related to the position so seems to be due to gel bridge. As it often happen, we want to replace the DRL/CMS 'pin' cable with the new set with one 'pin' electrode (DRL) and a flat electrode (DRL).
My question is if there is a recommend position for these electrodes. We want to measure somatosensory evoked potential so we are interested in the signal from the central position (FC and C channels).

Thanks,
Giuliano

Re: DRL gel leak

Posted: Fri Mar 10, 2017 12:17 pm
by Coen
Use less gel.

The question about the DRL and CMS position has been beaten to death on this forum.

Re: DRL gel leak

Posted: Mon Mar 13, 2017 7:58 pm
by Giuliano
The problem is that in our case the subject grasps two braille stimulator (with two different stimulation pattern) and probably a small amount of current flow from these devices to the DRL electrode. So this is not only a problem of power-line noise, but in P2 it is clearly visible the same pattern of the stimulation that we assume is due to gel bridge. However moving the DRL on the mastoid to avoid this gel bridge we suspect there could be some problem due to the asymmetry respect to the two different current flowing through the arms.
I know that the DRL drive the subject potential to the same potential of the AD-box in order to reject the common mode voltage and I understand the explanation in the article suggested in other post, but I don't know how to model this case in which the subject is touching a device connected directly to the power.

Re: DRL gel leak

Posted: Tue Mar 14, 2017 12:36 pm
by Coen
DRL is best located between the contact point of the interference source and the measuring electrodes. So, locate the DRL on the neck (leakage current flows from hands to neck, and does not cause potential differences between the EEG electrodes). Use a set with pin CMS (in headcap) and flat DRL (on neck).

Best regards Coen (BioSemi)

Re: DRL gel leak

Posted: Tue Mar 14, 2017 5:07 pm
by Giuliano
Thank you for the suggestion. Tomorrow we will try with this configuration.
In any case I was wondering if there is some technical specification about the circuitry of the amplifier. Because I am trying to understand how it works but I find it very difficult. I have understood that each channel is amplified respect to the reference voltage of the AD-box, so I guess that for each channel there is a differential amplifier in a non-invertent configuration and the common voltage is subtracted only afterwards (so this is different from the classic monopolar configuration). If I consider the CMS like a ground electrode as it is suggested in another topic of this forum I understand that the ddp between this electrode and the ground of the ad-box is the common mode voltage, but in general the ground should be positioned in an inactive place, and this is not the case, so on the CMS we can find both the common mode and also a potential due to the brain activity. This voltage is than used to drive the subject potential as close as possible to the ad-box reference through the right leg system. However the CMS is also used as 'digital' reference for the other electrode so I think in the raw data all channel can be considered as referred to the Cz electrode (not exactly Cz but the CMS is very close to Cz). Is that right?

Re: DRL gel leak

Posted: Tue Mar 14, 2017 5:27 pm
by Coen
There are no active or inactive places, there are only potential differences (voltages). The raw data is with respect to a point between CMS are DRL, depending on frequency (as discussed endlessly on this forum)

Best regards, Coen (BioSemi)

Re: DRL gel leak

Posted: Tue Mar 14, 2017 7:53 pm
by Giuliano
Sorry if I insist, I have read the other topic but there is still something not very clear to me: All the amplifier system that I have seen are based on an instrumental amplifier made by the classical configuration of 3 operational amplifier in which the input of the first stage are two signal from the body, so we have a differential signal (the voltage) that is our signal of interest and a common mode voltage, that usually drive the subject potential through the right leg circuitry.
In this case we don't have a real reference, so I guess that there is only one input to the operational amplifier, so I don't understand if it is used the same configuration with just an input from the body and the other input to the reference of the AD-box or something different.
Moreover I understand that the CMS is made like the other sensing electrodes, so theoretically it is just another channel but it is used only for the DRL.
Maybe I am going too deep in understanding how this system works but I am (almost) an engineer and I need to justify the choice that I make.
I appreciate your willingness.

Re: DRL gel leak

Posted: Wed Mar 15, 2017 5:42 pm
by Coen
The classic instrumentation amplifier with Ground or DRL (3 electrode measurement) is the standard solution for a single differential measurement (measure a potential difference with rejection of Common Mode). When you expand this design to more channels, several setups are possible, each with its pros and cons. BioSemi designed the so called "zero-reference" setup 20 years ago, see http://www.biosemi.com/pics/zero_ref1_big.gif. The reasoning underlying this design is discussed in viewtopic.php?t=33 and viewtopic.php?t=177 . The design was made possible by the excellent linearity of modern ADCs (Common Mode is converted to digital, and rejected by subtraction of digital codes in the computer) and has major advantages with respect to alternative setups in terms of channel separation (a disconnected electrode does not affect other channels), symmetry (all inputs have equal impedance, all analog outputs have equal load), and stability (all amplifiers drive a single ADC input).

Best regards, Coen (BioSemi)