Hello,
This is a beginner’s question, but I just wanted to clarify the steps involved in the very simple multi-level, multi-session analysis, similar to the one described in the FEAT 2 practical;
http://fsl.fmrib.ox.ac.uk/fslcourse/lectures/practicals/feat2/
where you have;
1)First-level analysis for each run
2)FE second-level analysis across runs
3)ME third-level analysis for between-subject analysis
Because I have two runs for most subjects, but have only one run for some, I ran a single second-level FE analysis, setting contrasts to calculate the mean across runs for each subject, as in the practical.
The practical then instructs you to choose “Inputs are 3D cope images”, and then enter cope images corresponding to each subject (corresponding to stats/cope1.nii.gz to cope10.nii.gz, for subj 1 to 10) for a specific contrast (lev2.gfeat/cope1.feat in the example). Does it mean that this analysis should be repeated for all the contrasts of interest, if you have more than one contrast?
Also, I’ve seen other instructions/ pipelines running FE second-level analyses individually for each subject to get the mean across run for the subject (creating gfeat folder for each subject), rather than doing it within a single analysis. Would there be any difference between the two approaches, beyond how you select the input images for the third-level analysis?
I know people have asked related questions, but I just wanted to make sure I understand the best approach.
Thank you in advance for your help!
Ami Tsuchida
|