Noise in Cyton data using Matlab

ahsanic · September 2017

Hey,

I have plotted the openbci sample data in matlab and here are the results.

When I acquired my own data from cyton board and plotted it in matlab, the results were like this:

Anybody knows what I did wrong, when I acquired the data, the electrodes where properly intact, I used two gels (ten20 electrode paste & neu prep gel). Why there is still noise in my data then?

Also comment on the y-axis limit of both data?

Thanks in advance.

wjcroft · September 2017

What is the scale used in the Matlab plots above? What does the Cyton data look like with the OpenBCI_GUI?

ahsanic · September 2017

http://www.filehosting.org/file/details/695919/3_min_eye_closed.txt

This is my data. 3mins eye closed.

ahsanic · September 2017

https://ibb.co/i7iOuQ

this is the matlab code we are using.

wjcroft · September 2017

You didnt answer my question directly in the thread here. Is the scale on your graphs in microvolts? There is some naturally occurring noise in all A/D systems. See this post,

http://eeghacker.blogspot.com/2013/12/self-noise-of-openbci-with-more-data.html

http://eeghacker.blogspot.com/2013/12/self-noise-of-openbci.html

I'm not going to give my email address to a file posting service. Please post anything you want to upload to a friendly service such as Google Drive or Dropbox. Again, "what does the Cyton data look like with the GUI"? I'm talking about the graphical output, not the raw CSV data. The raw data will have a millivolt level DC offset component that must be removed with a highpass filter, typically placed at around .5 hz. The GUI graphs have already done this filtering.

ahsanic · September 2017

Appreciate your help. Your knowledge regarding this is great!

Here you go: https://1drv.ms/t/s!AnQQsY6vhRLmg0OCv2Q_dHM8tigI

Also kindly tell me that when I open the data provided by OpenBCI the values were in the following format:

-4.40, 2.12, -3.89, 3.35, -7.11, -5.23, 4.18, -4.27

But our values are in the following format:

143628.14, 148204.67, 139280.70, 141125.72, 150712.63, 142664.61, 93198.92, 140597.08

So do you think the data provided by OpenBCI is processed in Matlab to get it like this format?

Thankyou so much.

wjcroft · September 2017

Hi, on my previous post I asked you to post a screenshot of the OpenBCI_GUI EEG channel graphs. Much easier to look at. Your CSV text file link is normal. As I mentioned the DC millivolt offset must be removed in Matlab. I don't quite understand your question, possibly a language issue. I'm not a Matlab expert. There are some other threads regarding Matlab use of OpenBCI data, see,

https://www.google.com/search?as_q=matlab&as_sitesearch=openbci.com

wjcroft · September 2017

More background on the FbEEG, Full Band EEG used in Cyton.

http://openbci.com/forum/index.php?p=/discussion/201/large-millivolt-data-values-fbeeg-full-band-eeg

ahsanic · October 2017

I have applied notch filter using MATLAB, results are quite similar to the sample data.

Thanks

wjcroft · October 2017

Because the data is FbEEG, you must apply a HIGHPASS filter, typically at .5 hz to remove the DC offset (which is in the millivolt range, 1000's of microvolts.) The notch is also helpful to remove extra mains noise. But the hipass is essential if you want to see the actual microvolt conventional EEG signal in the 1-40 hz range.

The GUI applies both the lowpass and notch to the graphs you see on the GUI times series plots.

pushtheworld · October 2017

@ahsanic looks like your values from the GUI are in microvolts while the values you are getting in matlab are volts.

Noise in Cyton data using Matlab

Comments