Muse data in numbers



I am embarking on analyzing muse data on Matlab.
I recorded data through MuseDirect and then converted the .muse file to .mat file using muse-player.exe
when I open the data matrix in matlab I get weird numbers hovering around the 800 (presimably uV). and the the 4th and 5th column in the data are NaN. the numbers in data.drlref are in orders of 10^6.
Can you please elaborate on whether these numbers are correct and whether I am getting the raw data correctly and how can I also get the reference recordings?



Around 800 uV is correct.
NaN means you’ve got dropped packet data: see here.


Thank you Enigma for your reply.

I got it. So what I did was to take the average of channel 1 and 4 ( mastoid electrodes) and substract it from channel 2 and 3 (the frontal electrodes) to get an eeg signal with avergae mastoid reference. Sounds right?


No, that won’t work. Imagine you had values of 1,1,1,1 or 2,2,2,2, both would give you results of 0.

If you want to normalize the values, all you need to do is subtract the average over the time period, or just minus 800 for the quick and dirty version.

So for a single second at 256Hz, take the sum of all 256 values and divide it by 256. This will give you the average over the second (probably ~800uV). The to normalize each value, subtract this average from each individual value.

So if the value T0 is 825, you’d get 25., if T1 is 750, you’d get -50. etc. etc.


Got it . Thank you very much


A followup;

I thought that the electrodes behind the ears at the mastoids are reference electrodes. now I see that the reference electrode is at Fpz. What you are suggesting is taking an average reference right? Does any other reference work?


Exactly. The signals you get through the API are already referenced, so you don’t need to do anything.



Is it possible to change the reference channel?



Not that I’m aware of. I think that would require reprogramming the firmware on the Muse.