A Theta Power Band Controlled 1D Cursor Task-Parameter Questions

Forum for discussion on different user applications
spine
Posts: 13
Joined: 02 Mar 2020, 22:07

Re: A Theta Power Band Controlled 1D Cursor Task-Parameter Questions

Post by spine » 04 Mar 2020, 21:43

Hello,

I meant to say that when I use Basket Task .prm directly and load it, I do not get the task shown in the wiki. Instead, I get a ball that starts in the middle of the screen that moves vertically (Which is good), but I then get a grey square that appears randomly across the screen too. But when I use the code that I last posted (Which was from this forum), we get the same task as the wiki.

I know you said to use the same parameters as in Basket_Task.prm, but which parameter file when I click CONFIG do I load to get the same task as in the wiki? Do I not load any parameter file at all and just copy the parameters from Basket_Task after clicking CONFIG? Does my question still make sense? What specifically do you mean by use Basket_Task.prm just to be clear?

Also, besides CAR, I know there are other options for spatial filters. What would, in your opinion, be a good place to start with the spatial filters besides CAR? Would laplacian be a good place to start (I am asking this because it seems like the newer version of BCI2000 has Laplacian added in the parms folder). I know that this is most likely something we will have to answer on our own, but I want to ask you first if you have any pointers where Laplacian fits into the picture.

Thank you!

Mohammad

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

Re: A Theta Power Band Controlled 1D Cursor Task-Parameter Questions

Post by mellinger » 05 Mar 2020, 07:04

I meant to say that when I use Basket Task .prm directly and load it, I do not get the task shown in the wiki. Instead, I get a ball that starts in the middle of the screen that moves vertically (Which is good), but I then get a grey square that appears randomly across the screen too. But when I use the code that I last posted (Which was from this forum), we get the same task as the wiki.
I just tested it, it works fine.

Make sure to proceed as follows:
1) Double-click on your self-created batch file to start CursorTask and the DSI source module.
2) In the operator module, click CONFIG.
3) In the config window, click CONFIGURE LOAD.
4) Make sure that each and every line there is checked.
5) Close the configure load window.
6) In the config window, click LOAD PARAMETERS.
7) Choose the file BCI2000/parms/examples/SMR_basket_task.prm and click Open.
8) Close the config window.
9) Click SET CONFIG.
10) Click START.
Do I not load any parameter file at all and just copy the parameters from Basket_Task after clicking CONFIG?
By "copy", do you mean that you manally enter the parameter values that you see in the prm file when opening it in a text editor? Don't do this. Even if you know the format of the parameter lines, you will be distracted by the overload of information, and inevitably make errors. Also, it is completely a waste of effort.
To load a parameter file, EITHER use the LOAD PARAMETERS button in the config window, OR add a line LOAD PARAMETERFILE to your batch file.
What would, in your opinion, be a good place to start with the spatial filters besides CAR? Would laplacian be a good place to start (I am asking this because it seems like the newer version of BCI2000 has Laplacian added in the parms folder).
A Laplacian filter is useful if you have localized sources, which you can make more sharp by applying a Laplacian filter. This fits the case of the mu rhythm. For the theta rhythm, you have to research for yourself.
My advice is to first stick with the CAR filter until you get ANY positive result AT ALL, and only then proceed to change anything in your setup.

-Juergen

spine
Posts: 13
Joined: 02 Mar 2020, 22:07

Re: A Theta Power Band Controlled 1D Cursor Task-Parameter Questions

Post by spine » 05 Mar 2020, 22:27

Hello,

I followed those instructions but still am getting the grey square appearing at random places. But I got an error saying the number of input channels isn't equal to the number of columns in the spatial filter (So the program would not run. It would only run and get the grey square appearing at random places on the grid after each trial has finished if I use CAR to bypass the error). Do you think if I checked all of the parts of the program after clicking config and compare it to the basket task parameters it should be enough? For example, I went line by line in the GUI to see what was typed in AR Order, Buffer Length, etc and made sure everything matched up to the parameter file I posted and found that was posted by Peter. The program seemed to run normally.

Also, do you recommend a 0.1 Hz high pass filtering? The documentation references it but the basket task does not use any filtering of the data.

Finally, what would you suggest as a first step for a laplacian filter/ a better filter than CAR? I do not expect you to know too much about Theta, but just a starting point on what weights to use would really be helpful.

We are having some issues getting the cursor to be controlled and it is probably because we are not looking at the right electrode/frequency as you suggested in your first post.

Thank you,

Mohammad

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

Re: A Theta Power Band Controlled 1D Cursor Task-Parameter Questions

Post by mellinger » 06 Mar 2020, 09:24

For example, I went line by line in the GUI to see what was typed in AR Order, Buffer Length, etc and made sure everything matched up to the parameter file I posted and found that was posted by Peter. The program seemed to run normally.
Peter posted those parameters so you could paste them into an empty .prm file, and then load that. Also, they should be identical to the contents of SMR_basket_task.prm.
Also, do you recommend a 0.1 Hz high pass filtering? The documentation references it but the basket task does not use any filtering of the data.
Do you mean, the DSI source module documentation suggests it? Then use it.
We are having some issues getting the cursor to be controlled and it is probably because we are not looking at the right electrode/frequency as you suggested in your first post.
Do an initial session as I suggested earlier. Do some research about theta. At the very least, enter "theta wave eeg" into Google and look at the first few results.

-Juergen

spine
Posts: 13
Joined: 02 Mar 2020, 22:07

Re: A Theta Power Band Controlled 1D Cursor Task-Parameter Questions

Post by spine » 09 Mar 2020, 10:13

Thank you so much for all of your help Jürgen! If any more questions arise in the future, I will be sure to ask.

Post Reply

Who is online

Users browsing this forum: No registered users and 4 guests