Dear Expert
I want to show that my ROI interacts with some areas more in condition A than condition B (rather than just showing interaction is greater in condition A than baseline).
So according to this web page:
http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/PPIFAQ
I have to create 2 new regressors (A-B and A+B).
***Q1************************************************************
But it's not easy to do so when I have many subjects….
(or maybe you can suggest a program to do this? )
*****************************************************************
Previously, my method was:
EV1= condition A (PSY regressor)
EV2= condition B (PSY regressor)
EV3= time-course of the seed ROI (PHYS regressor)
EV4= condition A's PPI (so I just select "between EV1 and EV3")
EV5= condition B's PPI (so I just select "between EV2 and EV3")
and then I create a contrast with EV4-EV5
***Q2***********************************************************
Can I still use this method?
*****************************************************************
Also, I have patient and control group with time 1 and time 2 data, my plan is:
(1) With a pre-determined ROI_1, run PPI for time1 and time2 data within each subject
(so that each subject has 2 PPI results)
(2) Run 2 separate group analyses for time1 and time2 to find out the group different regions
(3) Among these group-different regions, select a region for the following analysis (termed ROI_2)
(4) fslmeants -m ROI_2.nii.gz -i time1.gfeat/copeXX.feat/filtered_func_data -o time1.txt
fslmeants -m ROI_2.nii.gz -i time2.gfeat/copeXX.feat/filtered_func_data -o time2.txt
(5) Each subject has one value in time1.txt and time2.txt. Enter these values in SPSS, and do a repeated-measure ANOVA.
***Q3************************************************************
Can I use this method?
*****************************************************************
Thanks so much!
|