On 06/25/2013 12:12 PM, Colin Reveley wrote:
Doug - thanks that sounds well worth trying.
the labels we have have arbitrary spatial resolution in the saggital plane, and 1mm gaps between saggital slices.
In general in this project we are using an isotropic resolution of 0.25mm.
To generate surfaces in freesurfer I have to downsample to 0.35mm (to fit into 256)
in applying the procedure you suggest
a) is it possible to use a volume with a dimension > 256?
yes, you just have to have an accurate registration
b) is it possible to use a non-isotropic resolution?
yes
playing around with non-isotropic resolutions, projection fractions and vote would yield different results.
c) can mri_aparc2aseg output a volume > 256?
I don't think so. But you can use mri_label2vol to map the 256 to your volume, though you will probably lose some resolution
0.35 isotropic will be ok. we'd prefer 0.25 if we can.
best
Colin
---------- Forwarded message ---------- From: Douglas Greve <greve@nmr.mgh.harvard.edu <mailto:greve@nmr.mgh.harvard.edu>> To: <freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu>> Cc: Date: Sun, 23 Jun 2013 20:03:09 -0700 Subject: Re: [Freesurfer] ribbon label volume mapping You could run mri_vol2surf at different projection fractions (and using nearest neighbor interp). This will create a set of files, one for each projection fraction. You can then run mri_concat with the --vote option to get you the most frequently occurring label. You could then use mris_seg2annot to create an annotation. If you want to go the other way, you will need to combine your labels into an annotation (mris_label2annot), then convert that into a volume with mri_aparc2aseg. doug On 6/22/13 10:08 AM, Colin Reveley wrote:If one has a volume of integers, each representing a cortical area, can one map that to the surface as a set of labels, in a smart way that e.g. takes the mode of the voxels that lie on the vector between equivalent white and pial nodes (ie uses curvature and thickness information, tacitly or explicitly)? the labels are not exactly on the ribbon. There are many voxels outside the ribbon that are non zero. Like caudate etc. and some voxels that are within the ribbon but are zero because the label data is was/is non-isotropic originally (which is the problem we want to solve in fact). that would be great. Now let's say you already have labels on a mesh. can you generate a ribbon volume of labels, i.e. the opposite operation to what I enquire about above (opposite but not inverse operation)? This is work in rhesus. So my pipeline is very barebones indeed, really just mri_normalize ... mri_tesselate mri_inflate/smooth/sphere mris_fix_topology <-noaseg -noaprc> mris_make_surfaces <-naseg -noaparc> I only have access to what is generated by that sequence or what I can subsequently generate. I can generate a freesurfer like ribbon with mris_fill. My understanding of labels and data mapping in general in freesurfer is very poor. Maybe adapting one of your matlab scripts might be an idea, if there is not another way? many thanks, Colin, Sussex _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer