Using Neurosky to build a basic scenario

[BCL12]

Hello everyone!
I am new to the community and am in need of assistance for my academic work. I am an undergraduate student studying Human-Computer Interaction at Hampshire College in Amherst, MA. I have finished all of the basic tutorials and understand how to design scenarios, but I am now at a crossroads.
I am using Neurosky Mindwave Mobile and am trying to build a scenario that runs off of blinks because of it only having a single electrode, but I have no idea where to begin. I think it has to do with using classifiers, but I had trouble understanding them. I was hoping someone could try to guide me in the right path and get me to building my first scenario with Neurosky. I was thinking of trying to build something that uses blinks to change a color display from black to white and vice versa.
Any help would be appreciated as this is my last semester and a large part of my final project I need to complete in order to graduate.
Thank you in advance for your help,
Brendan

[nfoy]

Hi Brendan,
You might want to have a look at OpenVibe's website that provides examples of scenarios using classifiers.
http://openvibe.inria.fr/documentation- ... umentation
http://openvibe.inria.fr/motor-imagery-bci/

Best regards

[BCL12]

Hi Brendan,
You might want to have a look at OpenVibe's website that provides examples of scenarios using classifiers.
http://openvibe.inria.fr/documentation- ... umentation
http://openvibe.inria.fr/motor-imagery-bci/

Best regards

So I messed around with that scenario and tried to set it up for blinks instead.
I set the Lua stimulator's parameters for First and Second class to both be for eye blinks and the wait for beep and wait for cue parameters to zero.
No cue other than the crosshair shows up unlike the example scenario where a left or right arrow showed up, which I figured is okay.
I recorded a session where I would blink multiple times while the crosshair was showing up.
I then took the data and selected the electrode channel of the mindwave mobile and epoched the data based on the stimulations I set previously and sent that through the feature aggregator to the trainer.
I took the training configuration file, (which only produced 50% accuracy according to the terminal, which I figured was okay as this was my first test run) and applied the file to the classifier processor in the online scenario.
I ran the online scenario and would blink how I was blinking during the training scenario when the crosshairs would appear. The visualization showed no change though.
I'm curious where I went wrong.
Is it possible that it is working but no visible changes are being sent to the visualizer?

Thank you again for any advice you can offer