bci2000.org BBS

We are currently starting to work with BCI2000, and are in the process of creating our own signalprocessing modules and user applications. One of our scientists launched the idea to elaborate on the 3D task, by adding a 'stop signal'. The movement of the cursor would be using the power in the mu/beta bands, and the cursor should be 'stopped' at a target by the elicitation of a P300.
My question to this group is: what would be the preferred route to do this, is it possible to attach multiple signalprocessing modules to a single data source (guess not) or should I create a 'hibrid' from the ARSignalProcessing and P3SignalProcessing modules?
Any pointers are welcome.

Furthermore we created a source module for the use of TMSi Refa/Porti devices. If there is any interest we are willing to share (allthough it is still work in progress).

Thanks in advance,
Mark M. Span,
University of Groningen, the Netherlands

Mark,


These are great comments. While a number of people have thought about a hybrid BCI that could do both mu and P3, I don't think anybody has actually done it.

We currently cannot attach multiple signal processing modules. You would have to create a hybrid module yourself.

However, the most important question is whether this can even work in the first place. Thus, before you implement any signal processing online, I would suggest to evaluate this offline. To do this, I would change the task so that it contains, for example, a symbol that randomly flashes start/stop or so, so that the user might pay attention to the stop. Whatever task design you might choose in the end, you can then ask users to control the cursor and to pay attention to the stop at particular times. Offline analysis can then determine whether or not you are getting a P300 in the first place, and whether discrimination is good enough to permit online use. For example, you might have to design the task such that you can average a number of evoked responses.

Does this make sense?

The Gerv

Mark,

We would certainly be interested in incorporating the source module for the TMSi Refa/Porti devices in the BCI2000 project. After reviewing the code, we may be able to provide you with a CVS account that enables writing in your own directory in the source module. This has the added benefit for you that we might be able to suggest improvements to your source module that might make it more robust/general. In addition, we would be happy to note the inclusion of your source module on the News page of the BCI2000 web site.


Can you please contact Juergen Mellinger, the lead engineer of the BCI2000 project, so that he may review the code? His email address is juergen.mellingerNOSPAM_ATuni-tuebingen.de.

Thank you very much,
The Gerv