distinguishably identifying signals

[jamesLocal]

Hello,

I need to know in which Box(in the designer) and Project(In the source) does Openvibe distinguishably identifies the Right and Left hand signals.

Just to make it more clear,

1.) In the OpenVibe designer which Box is responsible to distinguishably identify Right Left Up Down hand signals from the GDF file. ? (I thought it was the "Stimulation based epoching" box, but when i connected the "Generic stream reader" to the "Graz visualization" ONLY, the hand signals were displayed)

2.) In the Source What is the project/files that are responsible for detecting RIGHT LEFT UP DOWN signals from the GDF file?

Thanks

[lbonnet]

Hi JamesLocal,

I need to know in which Box(in the designer) and Project(In the source) does Openvibe distinguishably identifies the Right and Left hand signals.

There is no dedicated box for that kind of detection. You have to use a classifier such as LDA, trained on some data. Classifiers handle 2 or more classes. The LDA we provide uses only 2 classes (e.g. right/left).
Please look at the motor-imagery scenario given in share/openvibe-scenarios/bci. You will find the processing pipeline we use for detecting left and right hand movement, using pre-recorded data in a gdf file. The classifier processor box in this scenario uses a configuration file, that was written in a training scenario involving a Classifier Trainer box.

Your questions prove that you need to get more information on how you can classify signals, and on how a BCI works in general. For example, there is no such thing as "left, right, up and down". You can try to detect if the user is moving his right hand, his left hand, his feet, etc. And you can't classify efficiently more than 2-3 mental states in the same paradigm.

Building a BCI is not simple. It requires a lot of preparation. We can help you here if you want to use OpenViBE, but we won't explain what is a classifier, why and when you need it.

Regards,

Laurent

[yrenard]

Dear jamesLocal,

in addition to what Laurent said, I would like to point out the future bnci website which has valuable materials to start with when new comer in the field of Brain Computer Interfaces.

Happy reading
Kindest regards,
Yann