[EEG] Re: [Noisebridge-discuss] Anyone interested in building EEG controlled robots?
jean.rintoul at gmail.com
Mon May 4 16:12:32 PDT 2009
P300 speller is just one of the examples it comes with. You can code up your
own system and plug it into their framework, giving it plenty of room for
expansion into new areas. You can run your own bizarro filters/matrix
transformations through mex files and everything.
Now I have BCI2000 running on a windows instance on my laptop. Looks like
there is a lot to learn about how to use it and it's very configurable. It
can import files from eeglab too which is rather handy. It has quite a few
little mini programs including one which looks very similar to actiview(the
BioSemi DataAcquisition software) for viewing and recording data, and even
has a mini stimulis presentation program(akin to E-prime or presentation)
with good time syncing(important). It works with most currently available
eeg systems out there, including Modular EEG(which is OpenEEG) - which is
this. BCI2000 is already able to work with OpenEEG hardware. This seems
Mac and Linux ports are due out later this year.
This thing may be a good way to record tons of data to make intelligent
classifiers to do cool stuff and control robots with?
(I was also doing a quick review of BCI2000 for work so it seemed relevant
to post here as well)
On Mon, May 4, 2009 at 1:03 PM, Kelly <hurtstotouchfire at gmail.com> wrote:
> Yeeeaaahhhh... I'm not that into P300. I don't really think that standard
> ERP is the future of EEG. But we should definitely try that out. I'm
> certainly not one to scoff at a pre-built framework. Scaffolding, yes
> please. And possibly for realtime BCI controllers, P300 is the way to go. I
> would still enjoy talking about the robotics end.
> On May 4, 2009 11:19 AM, "Jean Rintoul" <jean.rintoul at gmail.com> wrote:
> If people were looking for a flexible software infrastructure to hook their
> applications into there is this BCI2000 thing. It comes with a P300 spelling
> application as a starter too. Is it flexible enough to hook up to the
> OpenEEG setup? I have no clue.
> BCI 2000 allows you to plug and lay little matlab scripts and run them in
> real-time so you could make new algorithms on the fly. The spelling
> application classically runs with BioSemi. It sounds like you can use
> different hardware with it too. I have no personal experience but it sounds
> pretty neat from here.
> Run your classification algorithms, then have a stable set of functions and
> SDK at the other end to map to your twitter apps? Ideally, you'd have
> something equivalent to the err 'Emokey'... which just patches the output of
> your brainwave detections to different key commands so it can hook into any
> pre-existing software. It's still more useful to have applications specially
> built to run with the working detections, as they will map and work more
> On Mon, May 4, 2009 at 10:58 AM, Michael Shiloh <
> michaelshiloh1010 at gmail.com> wrote: > > Forget ro...
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the EEG