search

pwittchen / neurosky-android-sdk

Android SDK for the NeuroSky MindWave Mobile Brainwave Sensing Headset
https://pwittchen.github.io/neurosky-android-sdk/docs
Apache License 2.0
50 stars 16 forks source link

I want to ask about value given for each brain waves band #83

Open pascalisnala opened 4 years ago

pascalisnala commented 4 years ago

Hello I'm currently using your work with neurosky mindwave mobile 2, so far its working fine.

I technically don't really understand about how the brain waves converted from raw data to each bands (alpha, beta, etc). But, when I use your work to collect brain waves data, some data show very high values, exceeding 16 million. These very high values also seem to have a pattern that is like 167xxxxx image (please note that this is not only happen to alpha high but also other bands)

On the other hand, when I try to drop these enormous values, the data that is formed makes more sense to me image

Is this something that really should have happened? or there was an abnormality that shouldn't have happened?

I don't know if I post this in the correct way and place, but I hope it is Thank you

pwittchen commented 4 years ago

Hi @pascalisnala,

Thank you for detailed information. I haven't tested this software on mindwave mobile 2, so it's good to hear that it's working there. I checked it on version 1 only. I'm not really sure about these high values and I haven't performed many experiments with different brainwaves. My SDK just wraps around official ThinkGear SDK from NeuroSky and I don't perform any additional transformations inside. By looking at your charts, I can suppose that these very high values are some kind of false measurements or errors because difference between these values and other values is very high. For me it makes sense to discard these very high values and analyze the rest, but It's just my guess. If you would like to get more information, you can try to reach NeuroSky support. If you get some interesting insights, I'd like to hear them too :).

Regards, Piotr

pascalisnala commented 4 years ago

Hi @pwittchen

Thank you for the response it was really quick. I think I will try to find the optimal way about how to process this data for now. In case I find something, I will inform you. Thanks again!

Regards, Nala