 |
 |
Dear All, Our aim is to extract the volume of gray matter in a defined AAL template area. We are using FSL-VBM ). We did the analysis from fslvbm_1 to randomise without problems, i.e., we get the files we are expecting that we will need from reading FSL webpages. We do, however, have some questions that we feel we need to consult the FSL experts here on the mailing list. In fslvbm_2_template, the website (http://www/fmrib.ox.ac.uk/fsl/fslvbmsays template_GM_4D will be generated. We got struc/template_4D_GM.nii instead. We make the assumption that the two files are the same. The website's Section "C - Creating the template: fslvbm_2_template" says "... The registered GM images and their mirror images are concatenated into a 4D image called "template_GM_4D" and averaged to create the study-specific GM template at 2x2x2mm3 resolution in standard space. ". Since we have 15 nii files in template_list we were expecting struc/template_4D_GM.nii to have dim[4]=30 (for registered GM images _and_ their mirror images) so we were a bit surprised to find that dim[4]=15. It suggests to us that the mirror images were not included. Are we correct? Also, can we assume that the order of the 3D Gray matter images in struc/template_4D_GM.nii is the same as template_list, or is it _after_ sorting the template_list alphabetically? And if template_list is not provided, the order will be the alphabetical order of all struc/*_struc.nii Is the registration of the candidate GM images to GM ICBM-152 template in struct/template_4D_GM.nii a linear or non-linear registration? We assumed that there is now a voxel-to-voxel correspondence between the standard AAL template and each of the 3D volume in template_GM_4D.nii that we can simply overlay the AAL template on to the 3D volume and read off the value of voxels in a particular AAL region. Am I correct? In fslvbm_3_proc. all subjects in the study will be non-linearly registered to the study-specific template struc/template_GM.nii and concatenated into stats/GM_merg.nii and then modulated to stats/GM_mod_merg.nii. The website says the order of the 3D volumes in GM_mod_merg.nii (and GM_merg.nii) the same as the alphabetic order of files in the analysis directory when we call fslvbm_1_bet. That is good. As I want to find the average of grey matter probability for each subject, am I right to say that there is still a one-to-one voxel correspondence between AAL template and the individual 3D images in GM_mod_merg*.nii and GM_merg.nii that I can simply overlay the AAL template on to the 3D images and read off the values? Finally, as we have GM_mod_merg.nii and smoothed versions of it, GM_mod_merg_s[234].nii, is there any preference on which one I should use to extract the grey matter probability from? Many thanks in advance and Best Regards, Cinly