 |
 |
In case anyone is curious, it uses about 60GB to run it with 400 target parcels. Not too bad. > On Aug 18, 2017, at 10:16 AM, Saad Jbabdi <[log in to unmask]> wrote: > > Hi > I am afraid the only way to do this is to use —otargetpaths, which internally has a whole fdt_paths volume per target mask, which is why it is memory intensive. You could try splitting your 400 targets into, say, 10 groups of 40? > > Cheers > Saad > > > >> On 18 Aug 2017, at 15:08, Maxwell Bertolero <[log in to unmask]> wrote: >> >> I see from another post that: >> >> "In matrix2, the output is a MxN matrix where M is the number of seed voxels, and N is the number of voxels in the target2 mask. >> Each entry in the matrix is the number of streamlines starting in a given seed voxel that enter a given voxel or the target2 mask.” >> >> So I don’t think this actually has the information I am looking for. I don’t think I was clear enough before. Is there a matrix or output where I can constrain this so that I have this information my seed parcel to each target parcel? In essence, what I want is a matrix, where the first column is a voxel coordinate, second is one of my target parcels (there are 400), and third is the number of tracts that goe from my seed parcel to that target parcel through that voxel from the first column. >> >> Thank you for all your help. >> >>> On Aug 18, 2017, at 4:33 AM, Saad Jbabdi <[log in to unmask]> wrote: >>> >>> If you use -omatrix2 with a whole brain mask as target2, then the resulting matrix will store one fdt_paths per seed as rows of the matrix. >>> >>> Cheers >>> Saad >>> >>>> On 17 Aug 2017, at 21:39, Max Bertolero <[log in to unmask]> wrote: >>>> >>>> Is there a way to get a separate fdt_paths.nii image for each seed, so that, for each seed to seed, I know the most common voxels that the tracts went through? >>>> >>>> I am guessing that --otargetpaths might do this, but it seems to use a ton of memory to run. >>>> >>>> Thanks! >>>> >>> >> >