Dear experts,

 

I’ve downloaded more than 100 exams from ADNI, only 1 exam per subject. For each exam, I have one 3T 3D T1 MRI scan and 1 FDG PET image. The 2 images were coregistered. I’ve used SPM12 segment with default settings on each MRI. I’ve then used DARTEL to create a study-specific template and then normalized everything to the MNI space using the dartel derived transformation. If I load all the swmc1 images in checkreg and play the movie they look exactly identical, as you’d expect after a perfect normalization.

If I load all the sw PET images, instead, it looks like something is not right. Like if the cerebellum is lower in some patients, if lateral boundaries in the coronal planes move from patient to patient and so on. However, I can’t put my finger on anything clearly wrong. If I take an individual PET image and check it with the TPM, everyone looks ok. If I superimpose every segmentation with the PET, they look fine (both in native space and in the MNI after the whole processing).

 

Any idea on what I could check?

Has anyone had similar experiences? I’m sorry for being vague, but I can’t clarify the issue any better.

 

Thanks in advance

 

Luca