I do indeed have a hypothesis. But I will do my best to simplify this:
I run DR twice.
One with the following design:
input1: 'group=1' 'EV1 =score'
I get the following output:
dr_stage1
dr_stage2_ic
dr_stage2_subject
dr_stage3
I am interested in one particular network (dr_stage3 tfce * tsts1 & tfce corrp) which shows me the correlation between network and scores.
I run flsplit on dr_stage2_subjects####_Z.nii.gz since I want to see each subject's individual (the network of interest) network and how it correlates with the score.
I get the following files: vol0000 etc....vol0018
when I look at the volumes (vol0000 or IC's) they do not show any correlation to the scores.
I would need to run fsl_glm with the regressors to get such a correlation for the IC of interest for each subject, correct? ideally I would like to get the same correlation ouput (similar to dr_stage_3) for each individual subject. But that is not what I get when I split the concatinated IC's
Just want to make sure I am not missing anything. For the 2nd DR:
This is what my model looks like:
Design: Input1 Grp1 EV1 (grp) EV2 (scores)
similar outputs:
with the group dr_stage3_ic#### I am interested in one network
grp output: dr_stage3_stats1 shows me the network
grp Output: dr_stage3_stats2 shows me the correlation with that network.
When I run fslsplit on the dr_stage2_subject###Z.nii.gz, I get the IC's (vol###). These volumes represent the networks similar to the 'output dr_stage3_stats1' BUT NOT 'dr_stage3_stats2'. For that I would need to run fsl_glm with the regressors of interest. Correct?
Just want to make sure there is nothing that I am missing or that could have done differently.
Thanks, Jasmin
|