OpenViBE forum

Hello.

I am making an attempt at detecting foot motor imagery via beta rebound detection using the tie fighter demo/spaceship and neurofeedback scenarios.

When I run the calibration scenario, I get a large value for the mean and variance (~150). This value should be much lower according to the work published here: http://blog.jfrey.info/2015/03/03/openb ... r-imagery/.

Now, when I attempt to use this value in the next scenario to control the spaceship, it completely crops out all of the incoming data. I am left with no input signal.

Could anyone help me troubleshoot this?

Thank you!

-Travis.

Hi Travis, these numbers may depend on the data. Is the data and everything else identical? You can use Simple DSP box to scale signal and matrix streams. Likely the thresholding parameters are also modifiable... Somewhere.

Cheers,
Jussi

Thank your for your continued help, Jussi.

From the signal displays put in place, it seems like the data is being properly read and fed into the line of signal processing boxes.

So, are you saying I could scale my signal down by some scalar factor (say 1/2 for example) before feeding it into the "x*x" Simple DSP box?

By "threshold parameters", are you referring to those used in detecting/processing the signal during the online session?

Thanks again!
-Travis

I vaguely remember the spaceship has a sensitivity control somewhere, as I updated the doc in September. See the bottom of it how to control the threshold.

http://openvibe.inria.fr/tie-fighter-demonstrator/

Or if there's really a crop somewhere in the pipeline, its parameters should be modifiable. You can always use signal display or some other display tool to find out whats the data scale (enable the left ruler either by param or by clicking the top bar of the display). In practice its not uncommon that users fiddle with such parameters more or less ad-hoc to calibrate.


Cheers,
Jussi