EEG channel selection

[Gagandeep kaur]

Hello Grev,

can you please tell me what are the minimum number of EEG channels required to extract brain signal
i have planned to use p300 paradigm for my bci system, and if application is specific, plus brain area is identified ,(in in my case its sensory motor area), then is it possible to use single channel eeg and extract the brain signal to be used for experiments?

regards
Gagan

[gschalk]

Gagan,

If you plan to use a p300 paradigm, you probably want to record in parietal and occipital areas, not over sensorymotor cortex.

In either case, you can get near optimal performance with 8 electrodes for the p300, and with 5 electrodes (center+4 large laplacian electrodes) for mu/beta rhythms.

One channel will be suboptimal for both cases.

Gerv

[Gagandeep kaur]

Thanks a lot Gerv,
i further need your guidance...
im pursuing thesis in
brain computer interface... using p300 paradigm... uptil now i have
studied AVR microcontroller (ATMEGA16)... its interfacings.... i have
also been working on medical amplifier for ecg-signal using three
electrodes...that's just single channel , i got a very noisy
output....
earlier i planned to use this same single channel system for eeg
too... but now as you said i will use 8 electrodes...

please let me know if im working in right direction and also let me
know what line of action should i follow to accomplish my target...
sincere regards,
gagan

[gschalk]

Gagan,

I am not exactly sure what to say. As I mentioned, you will probably not get very far with only one channel. If you were to do sensorimotor rhythms, I would probably suggest to either use C3 or C4 with Cz as a reference.

Gerv

[Gagandeep kaur]

gerv,
i think i confused you

my point is that, to make a p300 paradigm based bci system, how should i proceed??

and yes i will use 8 electrodes ...

i found this info in one of bci2000 forum

Hank Janpol



Joined: 14 Aug 2007
Posts: 4
Location: Hawaii
Posted: Fri Aug 17, 2007 3:54 Post subject: electrode placement

--------------------------------------------------------------------------------

For brain signals that are easiest for subjects to learn control (Mu, P300 or other).

gschalk



Joined: 28 Jan 2003
Posts: 274
Location: Albany, NY, USA
Posted: Fri Aug 17, 2007 5:29 Post subject: electrodes

--------------------------------------------------------------------------------

Hank,

For 8 channels, we found the best electrodes to be:

for mu: T7,Cz,T8,F3,F4,C3,C4,Pz

for P300: P3,P4,PO8,Fz,Cz,Pz,PO7,Oz

Best,
Gerv

so can i also take theses electrode positions to get optimal results in my system? (
for P300: P3,P4,PO8,Fz,Cz,Pz,PO7,Oz)

regards
gagan

[gschalk]

Gagan,

Yes, you are correct.

Gerv

[Gagandeep kaur]

hello Gerv ......

may u please let me know that can i get near optimal eeg readings through less than 8 electrodes??? if yes then what locations are most suitable??

regards
gagan

[gschalk]

Gagan,

I do not think I can answer this question. We found the 8-channel montages by doing research with many subjects and many channels and finding the best subset of eight. You would have to repeat this process to find a smaller subset.

Gerv

[Gagandeep kaur]

Thank you once again Gerv.

Gagan

[Gagandeep kaur]

Grev,

i was going through bci2000 tour there i came across eeg measurement setup page , there its mentioned

Make sure your amplifier is connected to your computer, drivers are installed according to the vendor's instruction, and the amplifier is switched on.

i couldnt find that how to connect amplifier to computer, what drivers are to be installed? please guide me...

thanks
gagan[/quote]

[gschalk]

Gagan,

In order to record EEG with BCI2000, you will need an amplifier and digitizer system. BCI2000 supports several such systems. They are all listed on the wiki and on the main BCI2000 home page. The wiki also has tutorials how to write support for a hardware system that is not currently supported. Thus, please check out the documentation for further information.

Gerv

[Gagandeep kaur]

Hello Grev,

i performed EEG for detection of P300... i applied a flash as visual stimuli and did'nt get any significant change in response to detect p300... where could i possibly be wrong. tel me how should i do this experiment to detect p300

regards
Gagan

[gschalk]

Gagan,

I am not sure what to say. We have a comprehensive tutorial on the web that gives you a procedure that we know produces a good P300. Without any knowledge of your setup, I cannot really give you any reasonable suggestions.

Gerv

[Gagandeep kaur]

Grev,

i have a 24 electrodes, EEG system from Medicaid Systems... a Local Manufacturer in India...
I have been using a Photic stimulation to evoke the p300.

the electrodes are silver chloride coated..

now tell me how should proceed to get p300...

gagan

[gschalk]

Gagan,

In your setup, a lot of components are different than standard components. You are using a different amplifier and what looks like a different stimulation protocol. Thus, even if everything else in the system is the same, there may be an issue with one of these two. For example, the BCI2000 module for your EEG system may not be written correctly; or the parameters of your photic stimulation may not be appropriate. On the side: with simple photic stimulation, you will not evoke a P300 potential. Photic stimulation will simply produce a visual evoked potential (VEP) over occipital areas and not a P300 potential over parietal areas.

In either case, I suggest that you begin with a configuration that is as standard as possible and then begin to troubleshoot your whole system. For example, I would begin to simply record EEG, make sure you see EMG when you clench your teeth (similar to what is shown in the introductory tutorials on the Wiki), and that there is visual alpha when you close your eyes. If this is OK, you know that your EEG recording is good. Assuming this is fine, I would proceed with the standard P300 tutorial on the Wiki. We know that this will produce good P300 potentials in the largest majority of the subjects. Then, the remaining issue could be your photic stimulation protocol. Because I do not know anything about what you do, I cannot really make any good suggestions here, and I do not think it is practical to troubleshoot something like this from the distance.

Gerv