Skip to content

Home

Jump to bottom
yajing2018 edited this page Mar 20, 2018 · 6 revisions

2018/03/20

  1. checking the data after eddy_cuda (mr_dwi_denoised_gibbs_preproc.mif.gz)
    • three subjects discard several last slices; 10633: parietal lobe has a little cut-last 5 slices are gone (mr_dwi_denoised_gibbs.mif.gz is good) 12804: parietal lobe hass a little cut-last 4 slices are gone (mr_dwi.mif.gz discards has a little cut) 9810: cerebellar has a little cut
    • some volumes of 15 subjects are bad. 10014: 12 10110: 15;17;26;27 12458: 0:9 13091: 13;19;20 13305: 4;7;10;11;12;13;14;18;19;20;22;23;26 13359: 4;9;25 13681: 17;21; total volume = 22; 15558: 14;17;19;20 16034: 27 16035: 23;24; 9420: whole volumes are bad 9421: 15;16;20;26;27 9640: 13;22;26;27 9702: 7;9;12;18;19;20;23;27 9810: 26
  2. using the mr_dwi_denoised_gibbs_preproc.mif.gz to do step 3 and 4 for all subjects.

code:

cat "dtinames.txt" | while read sub;do cd ${sub}

#step3:estimate a brain mask

#dwi2mask mr_dwi_denoised_gibbs_preproc.mif.gz -force dwi_mask.mif.gz

the obtained mask by using dwi2mask looks weird, then create mask as following:

dwiextract mr_dwi_denoised_gibbs_preproc.mif.gz - -bzero -nthreads 8 | mrmath - mean -force meanb0.mif.gz -axis 3 -nthreads 8
mrconvert meanb0.mif.gz meanb0.nii.gz
bet2 meanb0 meanb0_brain -m -f 0.1 
mrconvert meanb0_brain_mask.nii.gz dwi_mask.mif.gz -force

#step4: bias field correction

dwibiascorrect mr_dwi_denoised_gibbs_preproc.mif.gz mr_dwi_denoised_gibbs_preproc_bias.mif.gz -ants -mask dwi_mask.mif.gz -fslgrad /Users/standarduser/Desktop/FTOdata/dwi.bvec /Users/standarduser/Desktop/FTOdata/dwi.bval -nthreads 8 -force

done

  1. The step 3 and 4 have finished. But some subjects failed to run dwibiascorrect.

Clone this wiki locally