Dear all,
I have some questions regarding SPM12.
I have only one group of 16 MS patients with baseline and one year structural images (I don't have longitudinal data from healthy control). I need to see the GM volumetric changes in these patients after one year relative to baseline.
These are the steps I am following as per posts in this forum:
1) Run the pairwise longitudinal registration to generate the subject average and Jacobian difference (jd)
-I left all default options untouched. I did not change Time difference since it was varying for each subject (some 12 months, some 15m, ....) and I am considering it as a confound factor in my statistical design.
- I left the bias regularisation of 1000000.
2) Segment the subject average, generating c1, rc1, rc2
- with no more Bias regularisation
- chose Native + Dartel Imported
- Warped Tissue: Modulated
- Deformation Fields: Forward
3) Use ImCalc to compute c1.*jd
4) Run Dartel, aligning the rc1 and rc2 images from all subjects together
5) Use Normalise to MNI space
-Used the final 6th iteration template
-Used "many subjects"
-Used flow fields from each subject
- Using Preserve Amount
-Used c1.*jd images for each subject
6) Run stats
Here are my questions:
1- Are these steps right?
2- Do I need to do any image pre-processing before these steps? I have done the following:
- AC orientation, NU correction before lesion segmentation, lesion filling
3- Is it necessary to use brain extracted in all those steps?
4- Should images be in stereotaxic space before any of the above steps?
5- I left the bias correction ON in the first step to correct for intensity inhomogeneity as result of lesion filling. Is it sufficient?
6- I use randomise to do my stats, do I need to smooth my concatenated images again?
- Is it right to use Single T tests, corrected for age, gender and time interval?
Thanks for your response,
Rezwan
|