Hi Martin -
You should be fine running probtrackx separately for different sub-
sets of targets, as long as you indicate to find_the_biggest the name
of all your files from both (or more) analyses.
You can also run different parts of the seed mask as separate jobs and
sum the results, that is done internally in probtrackx as each voxel's
results are summed up. (you can sum the fdt_paths files as well as the
seeds_to_blah files).
Cheers,
Saad.
On 3 Nov 2008, at 12:01, Martin Kavec wrote:
> Hi,
>
> I have a rather big probtrack/classification job involving 34
> anatomical
> targets in non-diffusion space. I tried to run the job on a QuadCore
> Intel
> 2.4GHz with 4GB of RAM, but the job didn't even start, because it
> consumed
> all the memory. I tried to cut the number of targets by half, which
> went
> fine, but after 7 hours of analysis, only 1/3 of the seed was
> classified. So
> my questions are following:
>
> 1. Can I concatenate the results, if I first run the classification
> of the
> seed to 1st half of targets and than to the 2nd half of the targets.
> I mean,
> would the find_the_biggest give me the same results, as if I would
> run the
> whole job at once?
>
> 2. I do not have so much machine time on the computer, so I was
> wondering,
> whether i could split the seed mask to multiple volumes each
> containing only
> one slice from the original seed and at the end sum up the results?
>
> Thanks in advance,
>
> Martin
>
Saad Jbabdi
Oxford University FMRIB Centre
JR Hospital, Headington, OX3 9DU, UK
+44 (0) 1865 222545 (fax 717)
www.fmrib.ox.ac.uk/~saad
|