So, depending on how you do this, you may need to rerun the tractography, because generating a CIFTI file that matches a particular order of voxels and vertices is not as easy as making a CIFTI file with the "usual" ordering of structures, and then taking the list of vertices and voxels from it to generate the seed files for tractography.
If you have just one consolidated volume ROI, you could use -volume-label-import and give the roi some arbitrary structure name (for instance, OTHER_GREY_MATTER), and use that. If you have separate volume ROIs for each subcortical structure, you can put them all together with -volume-math, then use the appropriate structure names from the -cifti-create-dense-scalar help, however this will put the structures in alphabetical order in the CIFTI mapping, so you might need to rerun the tractography with the new ordering of voxels. Similarly, -cifti-create-dense-scalar will put surface structures first. If you need to get around this, you should make separate CIFTI files per-structure as needed, and put them together in the desired order with -cifti-merge-dense. Tim On Thu, May 28, 2015 at 5:07 PM, Kristian M. Eschenburg <[email protected]> wrote: > Thanks for the quick reply Tim. > > After converting the correct metric files, how can I generate a customized > CIFTI file containing only the left subcortical ROIs and left cortex? To > clarify, I generated the left hemisphere subcortical ROIs from the > aparc+aseg.nii.gz volume in ${Subject}/T1w/. In my target list, I > specified the subcortical ROI volume first, followed by the left white > matter surface file. > > From looking through previous threads, it looks like wb_command > -cifti-create-dense-scalar might be what I need, but I need a volume label > file for the -volume <volume-data> and <label-volume> options. How can I > create a volume label file for only the left hemisphere subcortical ROIs? > > Thanks! > > kristian > > On Wed, May 27, 2015 at 6:13 PM, Timothy Coalson <[email protected]> wrote: > >> First off, this part: >> >> -row-surface fdt_paths_surface.func.nii -col-surface >> fdt_paths_surface.func.nii >> >> Those files should be named .func.gii, they are GIFTI metric files, not >> NIfTI volumes. >> >> Second, you appear to be doing tractography to subcortex and left >> hemisphere (only?) at once. In order to convert these, you will need to >> make a CIFTI subcortical and left cortex file with the same ordering of >> structures (and vertices/voxels within a structure), and use this with the >> -row-cifti and -col-cifti options instead of -row-surface and -col-surface. >> >> Tim >> >> >> On Wed, May 27, 2015 at 7:26 PM, Kristian M. Eschenburg <[email protected]> >> wrote: >> >>> Hi FSL >>> >>> I'm trying to generate connectivity matrices from the HCP data. Here is >>> my probtrackx2 call: >>> >>> SAMPLES=${HOME}Diffusion.bedpostX/merged >>> MASK=${HOME}Diffusion/nodif_brain_mask.nii.gz >>> SEED=${SurfDir}${SUBJ_NAME}.L.white.32k_fs_LR.surf.gii >>> WAYPOINTS=${HOME}ROIS/Left_WM_Bin.nii.gz >>> STOP=${SurfDir}${SUBJ_NAME}.L.pial.32k_fs_LR.surf.gii >>> AVOID=${HOME}ROIS/Total_Exclusion_Mask.nii.gz >>> TARG = list pointing to subcortical ROIs and >>> ${SurfDir}${SUBJ_NAME}.L.white.32k_fs_LR.surf.gii >>> SEEDREF = mask in seed space >>> XFM / INVXFM = linear transforms between T1 and DWI >>> TARGET2=${MASK} >>> >>> /usr/share/fsl/5.0/bin/probtrackx2 --samples=${SAMPLES} --mask=${MASK} >>> --seed=${SEED} --waypoints=${WAYPOINTS} --stop=${STOP} --avoid=${AVOID} >>> --targetmasks=${TARG} --distthresh=2 --nsamples=1000 --xfm=${XFM} >>> --invxfm=${INVXFM} --seedref=${SEEDREF} --dir=${DIR} --omatrix2 >>> --target2=${TARGET2} --opd --forcedir --pd --verbose=2 >>> >>> I am successfully able to run this call for both --omatrix2 and >>> --omatrix 4 (and by switching target2 and target4 masks). >>> >>> I convert the fdt_paths.nii volume using wb_command >>> -volume-to-surface-mapping <fdt_paths.nii> >>> <${SurfDir}${SUBJ_NAME}.L.white.32k_fs_LR.surf.gii> >>> <fdt_paths_surface.func.nii> -cubic. I can visualize this successfully >>> using Connectome Workbench on the L.white.32k_fs_LR.surf.gii. >>> >>> However, when I try to run wb_command -probtrackx-dot-convert using: >>> >>> wb_command -probtrackx-dot-convert fdt_matrix2.dot fdt_matrix2.dconn.nii >>> -row-surface fdt_paths_surface.func.nii -col-surface >>> fdt_paths_surface.func.nii, with out without the -transpose call, I am >>> getting this error: >>> >>> ERROR: found invalid index pair in dot file: 32472, 24335, perhaps you >>> need to use -transpose. >>> When adding the -transpose call, I receive: >>> >>> ERROR: found invalid index pair in dot file: 24335, 32472, perhaps you >>> need to remove -transpose. >>> >>> Do you have any suggestions as to what I'm doing wrong with this >>> specific call of -probtrackx-dot-convert? >>> >>> Likewise, when I try to convert the output from --omatrix4, from looking >>> at a previous thread by Dr. Glasser ( >>> https://www.mail-archive.com/[email protected]/msg00415.html), >>> I see that the -convert-matrix4-to-workbench-sparse call requires an >>> "orientation-file", which he calls >>> >>> >>> ${BedpostXFolder}/Whole_Brain_Trajectory_${DiffusionResolution}.fiberTEMP.nii >>> >>> >>> Where is this file? Looking through my bedpostX outputs, I don't see >>> any orientation files. >>> >>> Any help would be much appreciated! Thanks! >>> >>> >>> >>> Kristian >>> >>> _______________________________________________ >>> HCP-Users mailing list >>> [email protected] >>> http://lists.humanconnectome.org/mailman/listinfo/hcp-users >>> >> >> > _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
