noise, high impedance: electrode cap with Cyton+Daisy
Hello Everyone,
I am using OPENBCI 21 channel cap with Cyton and daisy board. My PHD thesis is to collect data for emotion classification using random visual stimuli. I have developed a MATLAB software to separate and label the collected data for classification. I have an issue that the data I am getting from my OPENBCI device is too noisy .I have settled everything in GUI (Smoothing On, Notch filter 50, Bandpass 5 to 50 Hz) All channels are NOT RAILED and impedance is Green with all channels though the values are high but channels are GREEN as per vendor recommendations.
Can anyone guide what to do .I am also attaching the images of my Eye closed of 2 minutes data. Application of filters is in MATLAB code is not creating any difference in it. I have read about cancel out the gain of the differential amplifier used in cyton board ((https://ieeexplore.ieee.org/abstract/document/9862906)but unable to figure out the gain of differential amplifier in my case as when i opened my EEG device in idle mode a file saved is this
Comments
Most EEG in the low microvolts range is noise. Your 2nd png file shows graphs with at most 1 uV of amplitude. Considering you also have pretty high impedances, I would caution making any conclusions from these very small amplitudes.
If you are using the EEG cap with proper gel (Electro-gel as sold by Electro-Cap or OpenBCI), then your impedances should be much better than shown. Is it possible that your reference connection is bad? Where / what are you using for reference?
Thank you so much wjcroft for your kind response.....My reference wire seems to be OK i am using the cap shown in the picture i am sharing ...It has a fixed GND and REF ,FZ,PZ,CZ pins..What should i do now Please guide
this cap"https://shop.openbci.com/products/openbci-eeg-electrocap"
Has the cap EVER worked as you expected? Have you emailed with customer service to ask specific question regarding your very low amplitude signals (1 uV)? (contact at openbci.com.)
Do the results look normal with the GUI, but extremely low with your Matlab? In that case, suspect your Matlab DSP operations.
One test that is typical with EEG, is to measure eyes closed alpha at a rear location such as O1 O2 Pz etc. For most subjects you will see decent amount of alpha production. Say around 10 uV.
https://docs.openbci.com/GettingStarted/Biosensing-Setups/EEGSetup/#4-alpha-brain-waves-eeg
Thank you so much @wjcroft ...
No When I try to plot .txt file which has been saved directly from OPENBCI the plots are similar as data taken using LSL MATLAB....
Stay blessed
What I'm asking is do you see normal eyes closed alpha with the GUI? And does that also appear with your Matlab?
Thank you so much @wjcroft
when i am recording my data i see very high amplitude in OPEN BCI GUI as i have shared a screenshot of that.......
Even when i am printing all channel data together in matlab i have a noisy plot like this
Your GUI snapshot shows that your signals are way too noisy. Please try to reduce your impedances. Suggest using an ohmmeter to check that your reference lead, connectors, etc., are not at fault.
As mentioned before your gold standard to try to achieve is the eyes closed alpha.
https://docs.openbci.com/GettingStarted/Biosensing-Setups/EEGSetup/#4-alpha-brain-waves-eeg
Can you explain how you are gelling the electrode sites? Perhaps that is also involved. Are you using the recommended Electro-Gel?
https://shop.openbci.com/products/electrodegel
Thanks @ wjcroft
Yes i am using recommended gel..... i bought with device
Referring back to your original image of the impedances, C3, C4, and F4 look decent, at 13 to 40 kOhms. However ALL the other sites are MANY hundreds of k ohms. So something is obviously wrong in your setup of the cap.
I would suggest just trying to gel a few sites in the rear, such as O1 or O2 or P3, and see if you get decent impedance and can replicate the eyes closed alpha. Maybe you are just not gelling sufficiently, or the gel is drying out, etc. Gel stops working when it dries out. Turn off channels not in use in the GUI by toggling the channel numbers on the time series.
thank you very [email protected] did and got same noisy data but now i will put more focus in gelling and reducing k ohms...
@wjcroft .....Do I need to work on differential gain which needs to be subtracted.
Thankyou soo much wjcroft......I will try ad will get back soon....
Stay blessed
Hello every one,
I had a long time noisy data issue from my OPEN BCI 16 channel cap ,Have tried so many things yet but nothing worked. Can anyone please tell me do I need to keep both SRB 1 and SRB2 "ON" in hardware setting as I am using an split cable to connect to REF to both (Cyton and daisy boards) well i just kept the default setting in which one SRB 2was ON..(https://openbci.com/forum/uploads/editor/kc/vpybpzwrvtt6.png "")
this setting
I kept this setting
Hello @Saima,
I merged your new thread "Recommended hardware setting for CYTON+Daisy board", into your already existing thread on the same subject (noise issues with your cap).
Just use the default setting that the GUI assumes for Cyton+Daisy usage. SRB2 is the default reference. SRB1 is NOT used. There is NO relation between your electrode cap noise levels, and the "hardware settings" in the GUI.
https://docs.openbci.com/GettingStarted/Boards/DaisyGS/
What happened with the previous suggestions given, regarding trying to achieve a reliable 'alpha' indication using 10-20 sites in the rear of the scalp?
William
hello wjcroft.....
Thank you so much for your kind gesture.......I was thinking to follow your given suggestion after making sure that everything is OK ,including Openbci GUI setting too thus no to leave any lacuna. As I have collected data so many times and it didn't work so I am afraid that device may get expire even after providing any noiseless data.
I will try to do it as soon as possible and will get back to you on this at earliest....
hello wjcroft
I am very thankful for your support....I collected data while gelling one two electrodes i.e. O1 and O2 with REF and GND and FZ,PZ,CZ .......By lot of effort I got impedance as shown in shared image.... I am also sharing OPENBCI TXT file of eye closed data. Can you please guide ....Is it seem to be better
~~
hello wjcroft
I am very thankful for your support....I collected data while gelling one two electrodes i.e. O1 and O2 with REF and GND and FZ,PZ,CZ .......By lot of effort I got impedance as shown in shared image.... I am also sharing OPENBCI TXT file of eye closed data. Can you please guide ....Is it seem to be better
~~
hello wjcroft
I am very thankful for your support....I collected data while gelling one two electrodes i.e. O1 and O2 with REF and GND and FZ,PZ,CZ .......By lot of effort I got impedance as shown in shared image.... I am also sharing OPENBCI TXT file of eye closed data. Can you please guide ....Is it seem to be better
~~
hello wjcroft
I am very thankful for your support....I collected data while gelling one two electrodes i.e. O1 and O2 with REF and GND and FZ,PZ,CZ .......By lot of effort I got impedance as shown in shared image.... I am also sharing OPENBCI TXT file of eye closed data. Can you please guide ....Is it seem to be better
~~
hello wjcroft
I am very thankful for your support....I collected data while gelling one two electrodes i.e. O1 and O2 with REF and GND and FZ,PZ,CZ .......By lot of effort I got impedance as shown in shared image.... I am also sharing OPENBCI TXT file of eye closed data. Can you please guide ....Is it seem to be better
~~
Hi,
Is there some problem with your computer or browser? You just posted FIVE identical comments??
I think you were trying post a photo / image. But on our Forum, these images must be limited in size, otherwise they are rejected. In other words you cannot post 10 megabyte images.
You do not have to 'PROVE' your impedance levels by posting an image. You can just state them in the text.
The larger point we are trying to make with your cap testing, is: can you get reliable eyes closed alpha increase. PLEASE answer: did that succeed or not?
William
hello @wjcroft
Thanks for your patience and support ...Yeah you are right I tried to attach the photos. I got the impedance of O2 around 60 and O1 55 ...but data i have recorded seems to be still very noisy and uncleanable. I got higher peaks at ALPHA 8 to 10 hertz but the highest peaks are between 20 to 25 Hz....
hello @wjcroft....
Greetings....I am rechecking my hardware connections too....Can you please guide cables of CZ,PZ,FZ remain unconnected...???
Any help will be highly appreciated
Saima