Dear SPM-experts,
I use SPM8 to do a VBM analysis. I am quite new to the VBM approach and have a conceptual question regarding the process with modulated and non-modulated images.
In my VBM analysis I want to include a comparison of the grey matter density as well grey matter volume between two groups. If I read the manual correctly, I have two options:
a) I produce modulated (mwc*, conveying volume information) as well as non-modulated (wc*, conveying information about relative concentration of grey matter) images using the 'New Segment' module. Those need to be smoothed and then, I can already do stats on them.
b) I can make the analysis using the DARTEL module. For grey matter analysis I use the rc1* files produced during 'New Segment' to create the custom template. Only in the ensuing normalization procedure I can then specify via the 'Preserve amount' (= modulation) or 'Preserve concentration' (= no modulation) option what I want. The result is then used for stats.
First of all: Are both procedures as I described them correct? I got a bit confused regarding the DARTEL procedure: If I produce modulated and non-modulated images using 'New Segment', can't I also use these files (wc* and mwc*) as the basis for DARTEL (i.e. producing separate templates for the modulated and the non-modulated case, and do normalization accordingly)?
Second of all: Is one 'better' than the other? I see that registration using DARTEL yields advantages as the custom template used for registration is based on the data I have such that group-specific anatomical variations will be accounted for. I did not quite understand where spatial normalization comes into play in the a) option: are the resulting images normalized and in MNI space (and if so, which template was used for registration)?
Thirdly: Is there any general opinion on what to use?
I'm very grateful for any suggestions that help me to get these things clear.
Best regards,
Ulrike
|