Ganglion Help needed, constant noise

shpSTEM

Hello OpenBCI and OpenBCI users,

We are STEM students that attend Seton Hall Prep, and we are attempting to use your product to gauge brain activity during various activities. These activities are assumed to reduce stress and increase relaxation. We are using the Ganglion and gold cup electrodes to see if this is the case. But we have encountered unforeseen issues that have impeded our progress.

While using the Ganglion and gold cup electrodes, we have only collected proper readings once. Even though we have replicated the process, our results get further and further from the expected readings. We followed the document "setting up for EEG https://docs.openbci.com/GettingStarted/Biosensing-Setups/EEGSetup/ " but this has not helped us in obtaining proper results. We have troubleshot using different electrodes, boards, and computers but the problems persist. The readings that we have gotten are abnormal and not consistent and we have been getting readings when the electrodes are not connected to a person. We could really use some guidance and any other assistance in setting up the Ganglion.

From,

SHP STEM team

Image below of "improper readings":

wjcroft

Hi SHP,

Has this Ganglion board EVER worked correctly when following the tutorials? When did you purchase? If this is a new purchase, and it never worked as expected, it is possible the board is defective. Your FFT shows a ~ 200 microvolt huge noise signal at 20 Hz. This is not normal, and does not even look like typical EMF (electromagnetic field) noise from power lines / supplies / lamps, etc.

You are using gold cups and paste? Locations? Did you check impedance?

William

shpSTEM

Hey William,
1) It has worked before and recently too. Or so we think, it aligned with what we commonly saw for reference online and certain stimuli made some things spike (not as drastic as the ones in the picture above).
2) We purchased it last year in 2020 around September/October, that is a rough estimate we will get you a more exact date.
3) We are using gold cups and paste
4) We followed the directions on the link we have attached for the locations. we used the pictures as a reference for where to place.
5) We have not really bothered with the impedance. Could you explain to us what that is and what we should be looking for?

We agree with you that the giant spike is not normal and we are trying to figure out what is causing such an anomaly. Another thing that might have not gotten mentioned is that when we take the wires off we will still get readings indefinitely, pretty much identical to readings as if the wires were attached. We thought that when we take off the wires everything would flatline, is that a fair assumption?
SHP

wjcroft

@shpSTEM said:
Hey William,
1) It has worked before and recently too. Or so we think, it aligned with what we commonly saw for reference online and certain stimuli made some things spike (not as drastic as the ones in the picture above).

The word 'spike' is not a good description of what your screenshot shows. What you are seeing is a pretty much CONSTANT 20 Hz sine or triangle wave noise source, with amplitude of around 200 microvolts. A more typical 'spike' would be like what happens with an eye blink. Spikes are short duration events. This noise is almost constant, with occasional shifts / changes happening at about 1 second intervals. The fact that the interruptions of the noise are happening at this 1 Hz interval is suspicious. Because the Ganglion compression algorithm also operates in this timeframe.

https://docs.openbci.com/Ganglion/GanglionDataFormat/#binary-format

I'm beginning to suspect that your Ganglion firmware has glitched. Can you try re-flashing it? Instructions are here:

https://docs.openbci.com/Ganglion/GanglionProgram/

Mentioning Richard @retiutut and @Shirley. However these instructions are fairly complex since they involve RE-BUILDING the Ganglion firmware from the source code. A much easier way to re-flash would be if OpenBCI staff provided the standard zip file used in the last OTA flash step, typically named something like: Ganglion_OTA_xxxx.zip. Richard or Shirley, can we give a Google Drive or other link to such a zip file? Another place to store it would be in the Arduino folder: https://openbci.com/arduino/

If the OTA zip is available, re-flash instructions become much simpler:

https://docs.openbci.com/Ganglion/GanglionProgram/#setup-mobile-device-for-ota-programming

2) We purchased it last year in 2020 around September/October, that is a rough estimate we will get you a more exact date.

That is good enough, so the board did work last year.

3) We are using gold cups and paste
4) We followed the directions on the link we have attached for the locations. we used the pictures as a reference for where to place.
5) We have not really bothered with the impedance. Could you explain to us what that is and what we should be looking for?

An impedance check with the GUI might add further info. But 'bad' impedance would not produce such a constant 20Hz noise.

We agree with you that the giant spike is not normal and we are trying to figure out what is causing such an anomaly. Another thing that might have not gotten mentioned is that when we take the wires off we will still get readings indefinitely, pretty much identical to readings as if the wires were attached. We thought that when we take off the wires everything would flatline, is that a fair assumption?

Well this behavior is further confirmation that you may have a firmware issue. However, EEG devices will misbehave when electrodes are not attached to the skin. But not with this type of constant 20 Hz noise.

Regards, William

shpSTEM

I apologize. I should have made it more clear for the first part. When I mention about the "spike", I meant that we saw consistent spikes on a time series plot not the FFT.

wjcroft

@shpSTEM said:
I apologize. I should have made it more clear for the first part. When I mention about the "spike", I meant that we saw consistent spikes on a time series plot not the FFT.

The 'spikes' in the Time Series graph in your screenshot, are not conventionally named spikes, but rather a long duration 20 Hz waveform, with occasional 1 Hz glitches. The 20 Hz time series does 'look' 'spikey', but is in fact is a very regular and continuous noise signal.

re: the OTA re-flash zip file, I am hoping Shirley can upload that for you from the lab.

wjcroft

SHP, hi.

Shirley is out of the office/lab today, but mentioned that she would look for the zip file tomorrow. She also suggested checking your per-channel switches on the Ganglion. These select whether the channel is operating differentially (such as for EMG / ECG), or with a common reference (as with EEG). If in the wrong position, one user in the past has seen similar noise.

https://docs.openbci.com/GettingStarted/Biosensing-Setups/EEGSetup/#1-connect-your-electrodes-to-openbci
https://docs.openbci.com/Ganglion/GanglionSpecs/#inverting-input-select-switches

William

Shirley

Hi SHP students,

We use a PCB scaffold to program ganglions. We don't use OTA. The user can follow the instructions to build from source https://docs.openbci.com/Ganglion/GanglionProgram/#building-from-source , and use the linked files here: https://github.com/OpenBCI/OpenBCI_Ganglion_Library/issues/17#issuecomment-470166588
Then follow the OTA file instructions: https://docs.openbci.com/Ganglion/GanglionProgram/#how-to-create-ota-file
Alternatively, since Seton hall is in t New Jersey, it would be a quick turn-around to re-program your ganglion board and send back.
The address is: OpenBCI, 67 West St, Suite 612, Brooklyn, New York, 11222
Should you have a question about shipping, please email [email protected].

best,
Shirley