Motor-Imagery BCI example issues

[ernestbofill]

Hi,
We are a group of master students from denmark getting started in the wold of BCI. We aim to make an experiment combining Virtual Reality and BCI. Introduction made, we are running in some trouble with the bundled MI-BCI scenario in openvibe.

1. In file "motor-imagery-bci-3-online.xml" says: The preprocessing of this scenario consists in re-referencing all channels to the channel 2 (which was located on the nose) and to select..." In our setup there is no electrode on the nose. On channel 2 we have FC3 then we have ground on the forehead and reference on the ear. To put it short, we don't understand what channel 2 and the nose have mean?

2. Then we get some errors when running that same file:

Code: Select all

[ ERROR ] At time 0.102 sec <Box algorithm::(0x000008cd, 0x000011f3) aka Reference Channel> Channel not found [Nz]
[WARNING] Box algorithm <Reference Channel> has been deactivated because process() function returned error

We believe this is related with the channel 2 thing I mentioned before, but not sure...

3. Do we need a particular positioning for the electrodes on the scalp? We followed the indications from this reference: viewtopic.php?f=17&t=399 c5, fc3, c3, cp3, c4, c6, cp4, fc4, ground on the forehead and reference on the ear.

Thanks for you help. I'm attaching some screenshots.
https://www.dropbox.com/s/mu21uzhr6b4i3 ... n.jpg?dl=0
https://www.dropbox.com/s/wayfbkgobrx6b ... n.jpg?dl=0

[fabien.lotte]

Hi,

I don't know which EEG system you are using, but I most current EEG systems the EEG channels are already referenced to the reference channel, so you typically don't need to use the reference channel box and can simply remove it (this box was used for an old EEG system in which the channel referencing was not done in hardware). For instance I use a g.tec gUSBAmp, which also has an ear reference, and I don't need to do anything in OpenViBE to take care of the reference, it's all done in hardware, and the EEG signals you read with openvibe are already correctly referenced. You should have a look at the motor-imagery-CSP scenarios, which are more recent, and don't use a reference box.

I hope this helps,

Best,
Fabien