Hi Pablo,
for the analyses you plan I would recommend to use the "MATLAB Toolbox
for functional connectivity (FC)" you can obtain here:
http://groups.google.com/group/fc-toolbox.
The toolbox provides the implementations of several measures of FC as
well as the option for a more sophisticated type of temporal smoothing.
Moreover, I would recommend to read the following paper that provides a
short introduction to FC and the toolbox:
http://www.ncbi.nlm.nih.gov/pubmed?term=zhou%2C%20siegle.
If you are only interested in the timecourse of the BOLD response of a
single condition you will have to extract the time course of the
region(s) (e.g. use Marsbar or the SPM VOI option) and read out only the
respective condition (and some postonset time window) by scripting which
is likely the easiest thing to do. If you have any further questions
don`t hesitate to ask.
Best regards
Raphael
Am 13.06.2011 16:38, schrieb Pablo Ripolles:
> Hello SPMrs,
> I want to perform a functional connectivity analysis on some fMRI data. We think that certain region A activates opposite to certain region B (When A goes up, B goes down and viceversa). We would like to use functional connectivity to see if both areas have a negative correlation.
>
> I have obtained the timecourses from the peak voxel at one specific contrast from both regions A and B. Then I have added them as regressors in the GLM. To see which areas have a negative correlation with the timecourse from area A I put a -1 in the contrast for this regresor. I do the same for the other area.
>
> We expect to see are A negatively correlated with the timeocurse of area B, and viceversa.
>
> I have some questions:
>
> 1.- First and most important, does this make any sense at all?
>
> 2.- If it indeed does, is it okay to put both timecourses as regressors in the same GLM or should I do this in two different GLM specifications?
>
> 3.- If I want to see the functional connectivity during one condition only and not the whole BOLD timecourse what should I do?Is it okay to enter as a regressor in the GLM only the values of the BOLD timecourse which happen during my condition and their respective scans?
>
> Thanks to all and sorry if my questions are very basic.
>
|