Dear expert,
Since I didn´t get a reply on my previous post, I will repost a shorten version of my question.
I want to run a VBM analysis in a group of stroke patients scanned at (t1: baseline, t2: after treatment 1, t3: after treatment 1 + treatment 2).
Based on the discussions on the SPM mail list, I followed these steps.
1. I used the “serial Longitudinal Registration”. I run 1 per subject.
In Time: I entered [0 0.15 0.19]. This corresponds with baseline (0), 8 weeks after baseline (0.15) and 10 weeks after baseline (0.19). Sound ok?
2. Segment the avg_*.nii images
I DID NOT RUN DARTEL. My reasoning was that since I´m working with lesioned brains DARTEL template may not be accurate. Is that right?
3. Segment the avg_*.nii images (mean of the 3 timepoints, 1 per subject)
3. Image calculator. c1avg .* j . Here I have a question. Since I have 3 j* images (from step 1), should I multiple the C1avg* for each of these j* images. So, the output of this step will be 1 image for each subject and timepoint. IS THIS CORRECT?
5. Normalise to MNI space . Used the forward deformation fields.
6. Run stats from the sw* images
¿At which step should I apply the smoothing?
Thanks in advance and best regards,
María
|