Hi Tim,
It's possible that you have run out of memory. Your data seem to be massively interpolated, that will increase memory requirements. The matrix size in the data file is much different than the scanning matrix you suggest. Apparently you have 480x480x48 voxels with size 0.52x0.52x3.26 mm, rather than 2x2x2 voxels.
Try running make_dyadic_vectors, in the bedpostx output directory as follows and see if that finishes:
make_dyadic_vectors merged_th1samples merged_ph1samples nodif_brain_mask dyads1
Have you performed any preprocessing before running bedpostx? Have you turned off scanner interpolation during acquisition?
Cheers,
Stam
On 31 Jan 2012, at 17:22, Tim van Hartevelt wrote:
> Dear FSL,
>
> We appear to experience some difficulties with running bedpostX on our DTI data.
> We have acquired our data on a clinical Philips Achieva 1.5 Tesla Magnet in Oxford. Diffusion MRI was acquired by using a single-shot echo planar imaging-based sequence with coverage of the whole brain; repetition time (TR), 9390ms; echo time (TE), 65ms. DTI images utilised 32 optimal nonlinear diffusion weighting directions (b = 1200 s/mm2) and 2 non-diffusion weighted volumes; reconstructed matrix=128x128x45; reconstructed voxel size 2.0mm x 2.0mm x 2.0mm.
> After a processing stage that takes up to a few days, we receive the following message (also see Processing stage of bedpostX below): For some reason the bedpostX process DOES NOT appear to have successfully completed. Please examine your results carefully.
> After careful examination we find that the dyads<i> files are missing as well as the mean_th, mean_ph and mean_f files. However, when running datacheck everything seems to be alright. We also find no problems with our dtifit. And to make matters more complicated, this problem does not occur with all of our datasets.
> Please find below the files after completing bedpostX, the output of datacheck, and the process log (that runs for several days instead of approx 20 hours).
>
> I hope you’ll be able to help us. Many thanks in advance.
>
> All very best,
> Tim
>
>
> Files present after bedpostX is done
> bvals
> bvecs
> commands.txt
> logs
> mean_dsamples.nii.gz
> merged_f1samples.nii.gz
> merged_f2samples.nii.gz
> merged_ph1samples.nii.gz
> merged_ph2samples.nii.gz
> merged_th1samples.nii.gz
> merged_th2samples.nii.gz
> monitor
> nodif_brain_mask.nii.gz
> xfms
>
> Processing stage of bedpostX
> Making bedpostx directory structure
> Queuing preprocessing stages
> Queuing parallel processing stage
> 0 slices processed
> 2 slices processed
> 3 slices processed
> 4 slices processed
> 4 slices processed
> 4 slices processed
> 5 slices processed
> 5 slices processed
> 5 slices processed
> 5 slices processed
> 6 slices processed
> 6 slices processed
> 6 slices processed
> …
> 45 slices processed
> 46 slices processed
> 46 slices processed
> 47 slices processed
> Queuing post processing stage
> 48 slices processed
> For some reason the bedpostX process DOES NOT appear
> to have successfully completed. Please examine your
> results carefully.
> Done!
>
> Data check
> bedpostx_datacheck forbedpost
> forbedpost/data
> data_type INT16
> dim1 480
> dim2 480
> dim3 48
> dim4 34
> datatype 4
> pixdim1 0.5208333135
> pixdim2 0.5208333135
> pixdim3 3.2567176819
> pixdim4 1.0000000000
> cal_max 0.0000
> cal_min 0.0000
> file_type NIFTI-1+
>
> forbedpost/nodif_brain_mask
> data_type FLOAT32
> dim1 480
> dim2 480
> dim3 48
> dim4 1
> datatype 16
> pixdim1 0.5208333135
> pixdim2 0.5208333135
> pixdim3 3.2567176819
> pixdim4 1.0000000000
> cal_max 0.0000
> cal_min 0.0000
> file_type NIFTI-1+
>
> num lines in forbedpost/bvals
> 1
> num words in forbedpost/bvals
> 34
> num lines in forbedpost/bvecs
> 3
> num words in forbedpost/bvecs
> 102
>
>
> --
> Tim van Hartevelt
>
> Department of Psychiatry
> Oxford University
> Warneford Hospital
> Oxford
> OX3 7JX
>
|