P300 accumulator- number of stimulations

About Brain Computer Interfaces, EEG, neurofeedback, related event announcements, jobs, etc.
Post Reply
etc
Posts: 23
Joined: Sat Mar 07, 2020 4:33 pm

P300 accumulator- number of stimulations

Post by etc »

Scenario BCI-P300 speller. New version 3.3.0, but it is the same with the old one.
I gave a look to the P300 Accumulator code to try to understand the reason of a problem(the problem was that the results, i.e. the characters selected, were not showed).
I read this note in the code:
“-- the idea is to push the flash states to the fifo, and when predictions arrive (with some delay), they are matched in oldest-first fashion.”.
Now, the number of the flash states is … the number of flashes.
But the number of predictions from the Classifier can be lower: if I have an Epoch block average=2, the number of predictions will be one half. If it is true, how can the P300 Classifier match 2 sets of informations where one of the 2 is one half of the other?
Thank you!

fabien.lotte
Posts: 112
Joined: Sun Mar 14, 2010 12:58 pm

Re: P300 accumulator- number of stimulations

Post by fabien.lotte »

Dear Ettore,

I was not involved in the coding of that specific scenario, so I am not sure I can help much, but I can still try :-).

The P300 accumulator accumulates the classifier output for each flash of the P300-speller, for each row or column, in order to find out which row and which column is the most likely to contain the P300 (and thus be the target row or column). But maybe you already know that, sorry if that is the case.
The Epoch average is set at 1 epoch, so normally the number of predictions should match the number of received stimulations. Probably the number of epoch averaged in that box should not be changed. (Again, I have not designed this scenario nor this code, so I am not sure why it is this way, but it may seem to me that this Epoch Average box is not really necessary, at least in its default setting :-)).

Regarding the error that you got, can you tell us a bit more? Did you get any warning or error message in the console, that could explain why the spelled letters are not displayed? Do you have the acquisition server well running (as the stimulations go through the acquisition server)?

All the best
Fabien

etc
Posts: 23
Joined: Sat Mar 07, 2020 4:33 pm

Re: P300 accumulator- number of stimulations

Post by etc »

Dear Fabien,
thank you for your answer!
1) about the p300 accumulator, I am sure that if I indicate in the box Epoch Average the parameter Epoch Block Average with Epoch Count 2 or greater, the P300 accumulator doesn’t work. Because the number of epochs is less than the number of flashes and the Accumulator can’t work.
If I indicate in the Epoch average box the parameter Moving Epoch average (immediate) with Epoch count 2 or greater, the accumulator works well! Indeed, in this case the average doesn’t reduce the number of epochs!
2) I attached the screen shot at the end of a run of the Replay scenario (P300-XDawn openvibe version 3.0.0). You can see that there is no result!
I had debug= true and it is possible to see that the accumulator choose the row and the column but the box P300 Visualization doent't show any result! Now I am working with
2020-12-11-p300.jpg
2020-12-11-p300.jpg (169.85 KiB) Viewed 594 times
the old version 2.2.0 with no problems on the same scenario and the same file.ov in input (i.e. the Acquision Server worked well).
ps I found a little bug in the version 3.0.0: in the scenario XDawn training, the box doesn’t update the file spatial.filter.cfg, also If I indicate with the flag my will in the parameters. (with the same parameters, in 2.2.0 the box updates the spatial.filter.cfg. )
Thank you very much!
Ettore

Post Reply