Dear SPM experts:
I am doing asymmetric longitudinal VBM analysis in stroke patient,the asymmetry part is following [A 12-step user guide for analyzing voxel-wise gray matter asymmetries in (SPM)],which I have no problem,but I am having some problems with the longitudinal VBM pipeline. Hope someone can help me.
The longitudinal VBM pipeline I read from previous posts is:
step 1 Serial longitudinal registration
step 2 segment the average template
step 3 run dartel(create template)
step 4 compute c1.(from segment) *jd (from step1)
step 5 normalize to MNI
Problems from step1
① serial longitudinal registration
times:Specify the times of the scans in years.
I am a little confused. I have three time point: first time, three months after the first time, a year after the first time, should it be [0 0.25 1]?
② Since my subjects are stroke patients,their brain can have quite big changes between three timepoints. (unilateral brain damage can deform the brain,but it can be recovered in the chronic stage),can I still use this longitudinal VBM pipeline to do the longitudinal VBM?(see attachments)
Problem from step3
① I read from the previous posts that lesion part can be set to 0 in the T1 and T2 so that it won't affect the segment. I notice that in the SPM12 6685,the c1 c2 rc1 rc2 will leave the lesion as a hole,but in the SPM12 7219 the hole will be filled with something from the TPM(I guess).
So I was wondering, when the hole is a hole, I can't use those rc1 rc2 to run dartel create template,instead I use the dartel template generated from the controls. but when the hole is filled, can I run the dartel create template' with those rc1 rc2?
Problems from step5
① When the c1.(from segment) *jd (from step1) is computed, do I have to do the modulation in Step5-Normalize to MNI ?
② the smw* from step5 is smoothed by default,but the input c* for step5 is filled with TPM,so I want to mask this TPM part before smoothing,but I don't know how,the SMOOTH-Implicit masking didn't seem to work for this.
Last question
If patients cann't do the run dartel create template',can I use my control subjects to create the dartel template and used in patients dartel procedure?
I have 30 controls,each was scaned twice,11 of 60 are of poor quality,so can I use the 49 to create dartel template?Or should I calculate the average template for each control and run dartel?
Looking forward to your reply.
best
wished
Yijun Chen
|