Hello SPM Experts,
I have a question about the iterative steps used during the preprocessing of my data. Specifically, after the program determines the normalisation parameters using the WM template (TPM), are the deformation parameters applied to the raw T1 MPRAGE and then registered to standard space before segmenting again and then smoothing. Again, I understand that these steps are not performed in sequence, but I am just curious as to whether or not two different registrations/normalisations occur (1 with GM TMP and 1 with MNI space)? Are there two normalisation steps?
Example:
*I know that the steps described below are incorporated into a single generative model, alternating among tissue classification (segmentation), bias correction, and normalisation steps, but I'm unsure as to where the deformation parameters are applied (to the original T1 before step 4 or not)? Any information on this would help. I have reviewed the 2005 Ashburner and Friston Neuroimage paper, which helped, but I want to make sure that I completely understand this process.
1. Segmentation and Extraction of a Structural Image (Raw Image) - the ‘native space’ images (raw image in figure) are segmented into separate GM and WM images, followed by removal of scalp tissue and other non-brain voxels.
2. Normalization of Grey and White Matter - brain extracted and segmented grey and white matter images from step one are spatially normalised. This is accomplished by transforming all of the subjects’ data into the same stereotactic space (white matter template).
3. Correcting for Volume Changes (Modulation)
4. Second Segmentation and Extraction of Normalised Brain Images - normalised structural images are now segmented into grey and white matter and a second extraction is performed to remove non-brain tissue.
Cheers!
|