Back to PLS HelpI'm Online
Posted on 09/02/14 12:07:55
Number of posts: 11
Dear PLS users,
I'm considering using PLS to obtain the correlations between brain data and behavior. In my case, the brain data consists on estimated sources from EEG scalp data, and the behavior data consists on a performance measure of a motor task and the experimental design (i.e. 3 conditions). I've been working on FieldTrip so far, so I have activation maps in structures compatible with MATLAB. I suppose that from this point, my problem is similar to the analysis for fMRI data, however I don't know if the PLS software is compatible with these data structures. If they're not compartible, what formats are? And has anybody encountered the same problem who could share with me the solution?
Best regards,
Laura
Replies:
I'm Online
Posted on 09/02/14 18:17:01
Number of posts: 229
Hi Laura..
Not quite sure what you have, but as long as the vector of data for each subject reflects the same spatial location and timepoint, etc, you can use PLS.. You will probably need to set up your own data matrices (datamats) and use the command-line version of PLS.. You would want one vector per person per condition, so a matrix with n(subj)x3(conditions) rows * y(datapoints->spatial location*timepoints) columns.
http://pls.rotman-baycrest.on.ca/UserGuide.htm#_Toc349137877
If you have a viewer for your input data, you should be able to port the output of the analysis into the same format for visualizing the results.
Nancy
I'm Online
Posted on 09/03/14 08:42:00
Number of posts: 11
Thank you, Nancy. I'll work on it.
Best regards,
Laura
Login to reply to this topic.
