PLS on time-frequency data
I'm Online
Maite Crespo
Posted on 06/21/10 08:17:41
Number of posts: 2
Hello!
I would like to apply PLS on time-frequency data of EEG activity.
1. How can I build up my data matrix, including not only the electrode and time samples, but also the frequency bins?
2. Is there any way of introducing this data to the PLSgui?
Thank you very much in advance.
Best regards,
Maité
Untitled Post
I'm Online
nlobaugh
Posted on 06/21/10 09:56:14
Number of posts: 229
Hi Maite..
you can run PLS on your time-freq data. To do this, you will need to use the command-line version. You would set up a datamat that contains bins within electrodes in a row for each subject/condition. The downside to this is that we have no easy way of displaying the results.
There is also a possibility of tricking the ERP module into loading the binned data - contact me offline, and we can see if it works.
cheers
Nancy
group analysis
I'm Online
Maite Crespo
Posted on 06/22/10 12:35:08
Number of posts: 2
Hello,
I applied the group analysis. The datamat_lst has the shape {2x1 cell} with 30 and 28 subjects in the first and the second groups, respectively. The output v is now a 4x4 matrix. How is the contrast for this analysis? How can I find out if there is an interaction effect condition by group?
The brainlv matrix has now 4 columns. Am I right considering the first column as the saliences for the contrast between condition in the first group, and the second column as the saliences for an equivalent contrast in the second group?
Thank you very much in advance!
Saludos,
Maité
Untitled Post
I'm Online
nlobaugh
Posted on 06/22/10 12:53:46
Number of posts: 229
Columns are the LVs
so column1==LV1, with contrasts between your two conditions in grp1 in rows 1-2, and in grp2 in rows 3-4
the significance of each LV is in your perm_result structure array: perm_result.sprob