P300 Speller and Neuroscan Synamp2

[azeater]

Hey everyone,

I've recently started using BCI2000 with a Neuroscan Synamp2 EEG system and have been running into some problems following the P300 Speller tutorial.

The operating system I'm using is Windows XP and I managed to follow the tutorial using the Signal Generator just fine, however, when using the NeuroscanClient, I run into some issues.

I modified the P300 speller batch file as suggested in the wiki, however get a "bad_alloc" error when I try to run it. However, using the BCI2000 launcher tool gets around this issue.

After I manage to get the experiment to start, the matrix screen does not run, it does not highlight rows or collumns and just stays still. I can still pause and resume the experiment, but no matter what the matrix doesn't begin highlighting.

I used neurogetparams to generate a prm file for the amplifier and load it as well as the P3InitialSession.prm file as per the tutorial. Everything works fine with the signal generator but it doesn't seem to like the Synamp2.

Any help would be really appreciated!

[boulay]

If you visualize the raw data, do you see reasonable data after you press set config (even before you press start)?

[azeater]

I must admit, I had not placed the electrodes on anyone's scalp because I just wanted to make sure that the experiment ran before going through all the trouble of setting someone up.

Is there some sort of threshold that must be met by the raw data before the experiment begins?

EDIT: So just to answer your question properly, the raw data was just small amounts of noise, no real EEG signal since the electrodes weren't attached to anyone's head.

[boulay]

I wasn't terribly specific when asking about the appearance of the data. Sorry about that. What I was looking for was to see if the data acquisition proceeded at a reasonable rate. That is, does the data visualization sweep through the entire length of the graph in 2 seconds? (2 seconds is the default) I'm not really concerned about the quality of the data at this point.

BCI2000 is designed such that the modules run independently of each other. If your system works with SignalGenerator, then replacing SignalGenerator with any other signal source module should still work, as long as that signal source is able to push data through the pipe at the correct rate. My intuition was that, because your problem was unique to the SynAmps module, and because the Preflight phase didn't throw any errors, that the system is stalled because the data are not coming through fast enough.

Using signal generator, can you go into the visualize tab of the config menu and check all of the boxes for visualizing the data then run the experiment? After you get a feel for how the data flow through the different filters, can you then try again with the SynAmp module to see if it seems correct?

Which version of BCI2000 are you using? If you downloaded from svn and built yourself then please provide the revision number. If you downloaded a release build then please indicate the release number.

Are you using Win XP 32-bit or 64-bit?

[azeater]

Sorry for the late reply, I only get a very limited time on the equipment due to other projects. I found out the problem and it turned out to be the sample block size was forcing the system to round down the stimulus time from the default 32.5ms to 0ms. After I set it to a multiple of 40ms, it began working.

Now I need to be able to pause and resume the P300 application from Matlab. I've managed to get all the UDP communication working and have managed to get it to stop with the "Running 0\n" command, but I can't resume it using this method. Besides modifying the source code with a custom "Pause" state, is there anything else I could possibly do to achieve this?