Dear DPARSFA experts, dear Chao-Gan
I just started using the DPABI toolbox (Btw, thanks for the excellent
software!) and ran into a problem when using DPARSF 3.1 to compute a
functional connectivity analysis. I pre-processed my resting state EPIs
using SPM5 (the data has been checked for quality and should be fine) and
then switched over to SPM8 to be able to use the DPABI toolbox. I have one
seed ROI (a 6mm sphere approximately covering the right ventral striatum;
MIN-coordinates: 9, 6, -8) and 56 subjects. During the FC-analysis
everything seems to be fine (at least I don’t get any error messages), but
when I compute a single T-test (using SPM8) the result looks suspicious,
there is one giant cluster covering the whole brain (k=49716), the peak is
located in the right ventral striatum (which corresponds to the centre of
the seed-ROI). I am not sure how I could resolve this issue, is this a
common or known error?
Thanks a lot!
Best regards,
Joe Simon
a network for supporting resting-state fMRI related studies.
Hi Joe,
Hi Joe,
Please read this thread: http://rfmri.org/comment/3622#comment-3622
Best,
Chao-Gan
Hi Chao-Gan
Hi Chao-Gan
Thanks a lot for your reply.
So if I understood correctly, you are recommeding to perform a "global signal regression"? And if I am not mistaken this means that you would include a "nuisance covariate regression" during pre-processing of the data? In DPARSFA there is an option to define "Gloabl signal" as a nuisance covariate, so this is probably what I should do to prevent the error?
Best regards,
Joe
Hi Joe,
Hi Joe,
Global Signal Regression (GSR) is still under debating, please see here: http://rfmri.org/GSRdiscussion
I may go without GSR and check GSR results in supplementary analysis.
Best,
Chao-Gan
Dear Chao-Gan
Dear Chao-Gan
I am still wondering what to do regarding my original question (the observation of a giant cluster covering the whole brain during the FC-analysis). I was not able to find a solution for this in the forum....
Best,
Joe
Hi Joe,
Hi Joe,