Two Channel Neurofeedback

[paulbrennan]

HI folks,

I have a few old Muse devices with two micro USB inputs. As I understand it, these can *both* be used to output EEG (though of lower quality, and only till the battery dies). I'd like to experiment with simple two channel neurofeedback with the external electrodes attached - if this is possible. The key is, I'd like to not be limited to protocols that use frontal montages of electrodes.

To do this, I would presumably need to do some signal processing of the Monitor output - presumably OSC - in some other software. I see there is software that is designed for this, like BioEra. Maybe that's overkill - could this be done in a very general signal processing environment like Max 4 Live? (This is the programming language that plugs into Ableton). The reason being that I've used it before, and also it works on the Mac and a lot of the more specialized solutions are Windows only.

Any other feedback on this plan would be welcome! It may be that the easiest thing is to switch the whole project to Windows.

Thanks, Paul

[James]

The Aux electrode(s) are only supported in the newer Muse models. Aux output on the old 2014 Muse with two usb ports is not supported by Mind Monitor, sorry.

[paulbrennan]

Oh I see, thanks for that James. I think there is some other way to get at that data, using older Muse tools, correct? Im sure those tools are unsupported but maybe they will still work. I see on the MAX forum that some people were able to do it somehow.

[James]

Yes, the old windows tools will work.
If I recall there was a bug in the SDK, where the wave values were calculated incorrectly when the aux electrodes were active. The 2014 Muse has the ability to output RAW EEG at 500Hz and I think the FFT math was expecting 256Hz and it screws up. But if you just want the RAW EEG, then you should be fine.
Also note that the 2014 Muse used a different data pin (2) than the later models (pin 4).

[paulbrennan]

Thanks for your help on that James. It might be over my pay grade but I will play around with it. Let's say I just use the 2016 - which I also have - then I only have one extra channel but for sure I can output it as OSC from MM. What do you think would be an easier signal processing chain to set up from there, if I just want to set up simple real-time sounds triggered by (e.g.) alpha ratios?

[James]

Check my Windows C# OSCMonitor receiver example here:
https://github.com/Enigma644

[paulbrennan]

Thank you

[paulbrennan]

I was able to quite easily download the old SDK for Mac and stream my 2014 headset data into MAX. Now I just have to parse the data and do something with it, but it seems quite possible.
One thing I notice though - there is no sign of the two extra electrodes in the EEG data stream and when I look at the SDK doc they are only mentioned once, and that is in the context of using them for *ECG* data (i.e. they are meant to be chest electrodes). In the MuseLab tutorial:

"In the screenshot below, the setup has the four EEG channels in the top graph. Channels (1) and (4) show three eye blinks. The middle plot shows the two AUX channels, ECG is visible from auxiliary chest electrodes. The bottom plot shows the three accelerometer axes."

Is it possible that the two aux electrodes on the 2014 don't output EEG at all?

[paulbrennan]

From FB group: "I don't know if this will be helpful but I used to work at Interaxon and if I remember correctly those were ports that allowed us to connect sensors to the chest or to the ears. They were custom build electrodes by our in house team."

[James]

You need to switch the preset to "AD"