Scaling BioSemi data

This forum deals with BCI2000 configuration issues.
Post Reply
inesplc
Posts: 16
Joined: 02 Mar 2016, 06:32

Scaling BioSemi data

Post by inesplc » 02 Jun 2016, 08:55

Hi all,

I'm looking for some clarification in the scaling of data acquired from BioSemi.
In here[1], the formula for calibration is:
CalibratedCh = (RawCh - SourceChOffset)*SourceChGain

My raw data is in digital units and I want them in uV [2]. Although default values for SourceChOffset are 0, with ActiView you can inspect this before acquisition and they around 20mV, or 20000uV.
SourceChGain is the inverse of the Least Significant Bit (LSB) [3] from the BioSemi amplifier, which from what I cant tell is LSB = 31.25 nV [4] and Gain = 0.0320nV or 0.000032uV.

Going with the above values in uV gives me EEG data under 1uV, which seems rather low [5]. Even if SourceChOffset is set to 0, the values still don't reach 1uV [6]. I've also noticed that during the recording with BCI2000 I mostly only see flat lines in the source visualization, not sure why.
Am I doing this correctly, or missing any steps? Shouldn't EEG values range from 1 to 100uV?

In addition, I couldn't find much information about the AUX sensors (gsr, plet) and the correspondent gain, but in my .bdf the max/min physical units are the same, so I assume the gain is equal.

[1] - http://www.bci2000.org/wiki/index.php/U ... urceChGain
[2] - http://postimg.org/image/uptusyz8r/
[3] - http://sccn.ucsd.edu/pipermail/eeglabli ... 01727.html
[4] - http://www.biosemi.com/faq/smallest_bitstep.htm
[5] - http://postimg.org/image/kd7k6w5pn/
[6] - http://postimg.org/image/4t06gcvl7/

pbrunner
Posts: 344
Joined: 17 Sep 2010, 12:43

Re: Scaling BioSemi data

Post by pbrunner » 02 Jun 2016, 10:43

Inês,

I assume that you made these plots in MATLAB. When you load the data into MATLAB using the load_bcidat mex function, you need to specify the '-calibrated' option to convert the signal from A/D units into uV:
load_bcidat

By default, signal data will be in raw A/D units, and will be represented by the smallest Matlab data type that accommodates them. To obtain signal data calibrated into physical units (microvolts), specify '-calibrated' as an option anywhere in the argument list.
As for the AUX, I would assume that the gain is the same. You can verify this by attaching a signal generator with a defined voltage.

Regards, Peter

inesplc
Posts: 16
Joined: 02 Mar 2016, 06:32

Re: Scaling BioSemi data

Post by inesplc » 02 Jun 2016, 10:55

Peter,

Didn't know that was an option, thanks! One question though: when setting the SourceChOffset and SourceChGain parameters, these values should be in uV, correct?
Also regarding the visualization problem: is the data that is shown in real-time calibrated or raw? If it's raw, than that might be the reason why I don't see it although it was saved and the system was working.

pbrunner
Posts: 344
Joined: 17 Sep 2010, 12:43

Re: Scaling BioSemi data

Post by pbrunner » 02 Jun 2016, 11:44

Inês,

looking at the source code there is only one allowed value for the gain (0.03125) which corresponds to how many uV the LSB represents. In the online visualization within the Operator you should see the signal in uV. You can right click on the display and enable show numbers to see the values.

Regards, Peter

inesplc
Posts: 16
Joined: 02 Mar 2016, 06:32

Re: Scaling BioSemi data

Post by inesplc » 09 Jun 2016, 08:25

Peter,

Just sent you an email with some data and print screens. I believe the might be some differences between the displayed and saved data.

Best,
Inês

Post Reply

Who is online

Users browsing this forum: No registered users and 28 guests