questions about timing parameters of P300 speller paradigm

This forum deals with BCI2000 configuration issues.
Locked
vattina
Posts: 2
Joined: 09 Jul 2007, 12:07

questions about timing parameters of P300 speller paradigm

Post by vattina » 09 Jul 2007, 12:11

Hi,

I am a new user of BCI2000, and I'd like to acquire new EEG data using the P300 Speller paradigm with the same parameters as for the BCI Competition 2003 (duration of intensification = 100 ms and duration in-between = 75 ms).
I have some questions concerning the parameters of this paradigm.
If I have well understood, the timing of the flashing depends on the "sampling rate" and the "sample block size" parameters of the source module. Do these parameters strictly depend on the hardware we are using? I am using a Brainamp EEG amplifier which has specific constraints regarding the samplig rate of acquisition (100Hz, 200Hz, 250Hz, 500Hz, 1000Hz, 2500Hz, 5000Hz but not 240Hz for eg) and the length of data block (10). Is it necessary that I put one of the sampling rate above and 10 for the sample block size in the rda source module?
How can I get a 100 ms duration for intensification and a 75 ms duration between two intensifications with a sampling rate of 250Hz and a sample block size of 10? Can I put float values for the "ontime" and "offtime" parameters in order to get the desired durations?
Can anyone help me please?

thanks in advance,

Virginie.

gschalk
Posts: 615
Joined: 28 Jan 2003, 12:37

P3Speller parameters

Post by gschalk » 09 Jul 2007, 17:15

Virginie,

As you correctly state, the timing in the BCI2000 system is dependent on the sampling rate and sample block size. For example, at 200 Hz sampling and a block size of 20 samples, the minimum timing resolution in the system is 100 ms. Thus, in this example, you will only be able to produce system timing (and thus, flashing times) that are a multiple of 100 ms. You cannot use fractions, e.g., an OnTime of 2.5. This decision, i.e., of using data acquisition to determine system timing, is deliberate as it allows for a good timing relationship between EEG samples and visual presentation.

In your example, at 250 Hz and sample block size of 10, you have a resolution of 40ms. Thus, you can do 120 ms on and 80 ms off, but not 100 and 75. Fortunately, the exact definition of these parameters does not matter much. If your stimulus update between 4-8 Hz, you should be in good shape.

Finally, if the Brain Amp only produces certain sampling rates and only allows for a sample block size of 10, then you need to use these parameters in BCI2000.

Gerv

vattina
Posts: 2
Joined: 09 Jul 2007, 12:07

Post by vattina » 10 Jul 2007, 13:21

Dear Gerv,

many thanks for your helpful answers.
I have few more questions about P300 speller paradigm and parameters.....
What is the usual size of the matrix display ? Is it full screen?
According to what I saw, it seems that there is a classification for each character, even if we are using the offline mode. I just want to record offline eeg data from the P300 speller paradigm of BCI2000 in order to test some algorithms afterwards. Are the recorded data preprocessed or are they "real" raw data (= the same as the ones delivered by the eeg amplifier)?
thanks in advance,

best,

Virginie.

gschalk
Posts: 615
Joined: 28 Jan 2003, 12:37

Re: P3Speller

Post by gschalk » 10 Jul 2007, 15:58

Virginie,

We use different sizes for the matrix. This includes full screen, but also only parts of the screen.

Correct, there is always a classification. There is a parameter ShowResults that you can use to suppress the output.

The recorded data are not preprocessed in any way. You can use the BCI2000toASCII program to convert data into ASCII/Matlab, or use the load_bcidat.dll mex file to load data into Matlab directly.

I hope this helps.

Gerv

Locked

Who is online

Users browsing this forum: No registered users and 2 guests