Large laplacian

Forum for discussion on different signal processing algorithms
Locked
aloplop
Posts: 41
Joined: 03 Sep 2008, 07:20

Large laplacian

Post by aloplop » 11 Nov 2008, 12:43

Hello,

I have performed new tests with 16 electrodes and a smaller EEG and I have found interesting results after the Stimulus Presentation Analysis. For example, after using just the left and right arrows and a deterministic mode I obtained r^2 ~0.34 with left and right arms imagery in the channel C4, which is best than I performed other times.

As it is said in another post the mu rhythm can change from one day to another so, as you have recommended to me I should use a large laplacian configuration centered in C4 (also I obtain good values of r^2 in P4).

My question is, after doing the Stiumulus Presentation and the OfflineAnalysis:

If I try the C4 Cz P4 T8 F4 configuration and I found that the r^2 peak is in the C4 channel I will use a Sparse Matrix for Spatial filtering similar as the one in http://www.bci2000.org/wiki/index.php/U ... tialFilter.

But if I find that the peak appears in the channel P4 what should I do?? Maybe using the CAR filtering ??

Thanks.

mellinger
Posts: 1208
Joined: 12 Feb 2003, 11:06

Post by mellinger » 12 Nov 2008, 06:31

Álvaro,

P4 is located over the visual cortex. It is likely that the large r^2 value there is due to modulation of the subject's alpha rhythm, caused by visual input during stimulus presentation, especially when the experiment was performed in a dark or dimly lit environment.

Thus, it is unlikely that P4 will be a good location for mu rhythm feedback.

HTH,
Juergen

aloplop
Posts: 41
Joined: 03 Sep 2008, 07:20

Post by aloplop » 12 Nov 2008, 07:51

Hello,

thanks for your indication. Finally, I decided to use the channel C4 as the center of the Large Laplacian and use the CAR filtering with the channel C4 as output, because if I introduce a matrix the same as in the tutorial for channel C4 using the Sparse matrix this error appears:

Code: Select all

SignalProcessing: caught exception EAccessViolation,
terminating module.

Thanks again,

Alvaro.

mellinger
Posts: 1208
Joined: 12 Feb 2003, 11:06

Post by mellinger » 17 Nov 2008, 09:03

Thanks for this bug report. I could not reproduce the bug with the current version of BCI2000, so it may be related to the issue fixed in revision 2167 (http://www.bci2000.org/tracproj/changeset/2167).

When you update to the latest version of BCI2000, do you still experience the described behavior?

Regards,
Juergen

aloplop
Posts: 41
Joined: 03 Sep 2008, 07:20

Post by aloplop » 18 Nov 2008, 11:07

Hi,

I updated this morning to the revision 2201. I forgot to try that with the gUSBamp. However, I have tried this combination:

Code: Select all

start operat.exe
start SignalGenerator.exe 127.0.0.1
start ARSignalProcessing.exe 127.0.0.1
start CursorTask.exe 127.0.0.1
cd ..
In the filtering tab I select Sparse Matrix and introduce as the matrix the rows from 6 to 10 of the example about the large Laplacian in the spatial filtering page: http://www.bci2000.org/wiki/index.php/U ... tialFilter, but the error still appears.

Thanks.

mellinger
Posts: 1208
Joined: 12 Feb 2003, 11:06

Post by mellinger » 19 Nov 2008, 08:17

Hi,

I don't experience a problem with the spatial filter here.

In the current version of BCI2000, the error message will be a bit more informative about the context in which the crash occurs. Could you post it?

Also, a parameter file would be helpful in reproducing the bug.

Thanks,
Juergen

fnijboer
Posts: 1
Joined: 07 May 2009, 05:42

Post by fnijboer » 07 May 2009, 05:52

aloplop wrote:Hello,

thanks for your indication. Finally, I decided to use the channel C4 as the center of the Large Laplacian and use the CAR filtering with the channel C4 as output, because if I introduce a matrix the same as in the tutorial for channel C4 using the Sparse matrix this error appears:

Code: Select all

SignalProcessing: caught exception EAccessViolation,
terminating module.

Thanks again,

Alvaro.
Hi Juergen,

I am getting the same error message as Alvaro. It sometimes happens during a P300 run. First I thought it only gave the error while it was doing the signalclassification (deriving the letter) but now I've also seen it happing while it was still presenting the flashes.

On the few occassions that it does not give this error message the system simply spits out a letter and everything works fine. I've changed the PostSequenceDuration such that it is long enough for the system to do it's calculation before the new target letter needs to be selected. I am using (I thik the latest) version of BCI2000 (v2.0).

I'll send you the screenshot of the error message and the prm per mail.
Thnx for your time.
Saludos desde España,
Femke
[/img]

mellinger
Posts: 1208
Joined: 12 Feb 2003, 11:06

Post by mellinger » 07 May 2009, 07:42

Hi Femke,

the error message is probably due to an invalid channel specification in the TargetERPChannel parameter on the Visualize tab. In your parameter file, you ask for visualization of channel 4 but there are only 3 channels present in the TransmitChList parameter.

I added a check for a valid TargetERPChannel, which should fix the problem.

Thanks for reporting,
Juergen

Locked

Who is online

Users browsing this forum: No registered users and 6 guests