Emotiv Neuroheadset

[George Avazzy]

Hello,

Take a look at the Emotiv EPOC, a commercial EEG used to control video
games. They're now shipping SDKs for developers and researchers. The
standard SDK connects to the 3 suites--cognitiv, affectiv, expressiv--
which detects various brain states.
They say to get raw EEG costs $25,000, but surely it would be possible
to develop a driver for free.

Link: http://emotiv.com/corporate/1_0/1_6.htm

Imagine the possibilities!

[T Biehn]

There is hard science to making the relation between measured
occurrence x and eeg behavior y.
The device spits out 'measurements', not actual EEG.

25,000 pays for the scientific research that spits out an EEG feed
(hard work.) The sub 500$ pays for the easy research of EEG -> Emotion
mapping. This does not mean the emotion drivers include a measurement
-> eeg -> emotion but rather they figured out eeg -> emotion then
replaced that with measurement -> emotion. This is the most efficient
way of doing it, and so you will not be able to easily duplicate the
EEG functionality.

-Travis

[Dennis Peterson]

I'm not sure I get why a raw EEG feed should be more expensive than a
processed EEG feed.

I used to know a physics student who built his own EEG. Opensource
designs if you'd like to try it are at
http://openeeg.sourceforge.net/doc/.

If you're interested in neurofeedback (alpha/theta waves, etc) there
are some pretty fancy devices available for a couple thousand dollars.

For this kind of work you're essentially looking for an analog-digital
converter, and some FFT to extract the amplitudes of particular
frequencies. A soundcard provides all the essentials, but
unfortunately filters out the low frequencies you'd be interested in.
Nevertheless there is an openeeg design that uses a soundcard and
works around this problem somehow.

[T Biehn]

Dennis,
I'm saying the charge more for the raw EEG feed because of the
development costs incurred in figuring out how to derive EEG from
headset metrics were likely high (original research) whereas the
research cost of EEG to emotion was probably inexpensive
(implementation of existing research).

They can't expect to collect on this debt off of the consumer use, but
those interested in creating EEG-derivative works will have to help
cover the cost of the research.

I'm sure their sales department would be more than happy to explain
exactly why they chose the cost model that they did.

-Travis

[Dennis Peterson]

Put electrodes on your head. Record the signals you get. That's all
there is to EEG. They say they're detecting electrical signals, so
there's nothing unusual here.

"Detecting various brain states" is where they derive information. The
headset metrics themselves are EEG. There's nothing to derive.

Business models are not necessarily based on technical merit. They
probably think that their headsets will become so popular that it will
be worthwhile for commercial developers to pay through the nose to get
full access to them. Given that EEGs are well-known technology, I
suspect that a more open competitor will eat their lunch, if this sort
of application catches on.

[T Biehn]

Got it.
I thought they were using some novel technique for collecting EEG data.

-Travis

[George Avazzy]

EEG is more useful than their presets, so that's why it costs more. But it's a USB device, so it should be possible to create drivers for it the same way that Linux developers develop drivers. http://www.linuxjournal.com/article/4786
What I noticed is that you pay $500 for the SDK+headset, but you don't get the headset for another 10 weeks.

But it would be interesting to do a study on the effects of quad n-back on EEG. At my university I participated in an EEG study in which they studied activity during visual 2-back and 3-back, but they only studied each participant once, never went to a higher n-back, and didn't include audio/shapes/etc. I wonder what the EEG data would look like during quad 6-back or higher.