Problems with blocksize and neuroscan

[Marco Rotonda]

Hi there,
I'm quite confused about the configuration of Neuroscan:
I got the params back from neurogetprams and there are 2 strange things:
1) AlignChannels = -1
2) SourceChTimeOffset = -1

Another thing. From the param file I've set the SampleBlobkSize at 40 (as should be in neuroscan).
I'm using FieldTrip as matlab buffer for signal analysis.
Either if I put AlignChannel at 0 or 1 and I try to read the samples from the buffer with matlab I have same strange jumps and stops like this:

4120
4160
4240 (here are 80)
...
4400
4520 (again)
4560
...
4720
4800 (again)
4840
4880
4920
5000 (again)
5080 (again)
...
8160
8200
8280 (again)
8280 (no new samples?)
...


I can not understad if these stops and jumps came from the acquire buffer or the matlab buffer or whatever, or it's just some misconfiguration...
Maybe there could be some problems running 2 server buffer on the same machine (one on port 3999 for neuroscan and one on port 1972 for fieldtrip)?

thanks in advance

Marco

[gschalk]

Marco,

Please configure the AlignChannels and SourceChTimeOffset as described here:
http://www.bci2000.org/wiki/index.php/U ... mentFilter

AlignChannels should be zeros and SourceChTimeOffset should be as many zeros as you have channels.

(neurogetparams was written for BCI2000 v1.0 and may not have been adapted for v2).

Then, I would use the neuroscan Source module in conjunction with existing modules, e.g., ARSignalProcessing and CursorTask. Once this is working, I would try the FieldTripBuffer. You have a complicated setup and it's difficult to debug as a whole.

Gerv