It's not an issue, but I need help

[iaaly]

Hello,

Sorry for submitting an issue but actually I didn't find a way to contact you but this.
I'm working on an SSVEP Project, your toolbox helped me alot, I did the acquisition and then offline processing and classification using this toolbox and obtained a great accuracy.

Now I'm looking forward to do online testing, but I don't have an idea from where to start..

I mean, where does exactly in the code the decision is taken to chose among the frequencies (labels) for a specific signal, and any ideas to do it live to a real time signal?

Thank you for your help

[liarosge]

Hi iaaly,

There is an example on how to do online testing in branch "lsl". See the file "exampleLSL.m" and "LSLWrapper.m", maybe you can adjust them to fit your needs. You have to provide a trained classifier for it to work, which is done by calling the method "build" in a classifier object after adding the instances.

It is based on the labstreaminglayer library (the lsl library is required to work), and reads 2 streams, a data stream and an event stream. When a specific event code is sent to the event stream, the classification pipeline reads the last 5 seconds of the EEG signal and outputs a result.