 |
 |
Dear John, > Dear Jasper, > I have DWI data (The images were acquired using a single-shot, spin echo, and echoplanar imaging sequence with twice refocused spin echo diffusion preparation (Q-ball imaging). Each series included 60 images acquired with diffusion weighting and 8 noncollinear directions. are you sure about this? 8 non-collinear directions sounds very little, and why would you not go for 60 non-colinear directions? > > I ran eddy_correct and I calculated motion parameters and SNR. Motion is excessive!! in this data set, and as a result we need to exclude large number of my subjects (~60%) of the data. I am hopping that using "topup and eddy" to correct motion in this data set would help a bit. I did the following: > > The raw dwi data is in the form of nifit "dwi.nii" is a longside with "bvals" and "bvecs" within a folder named "subj1" > > cd subj1 > fslroi dwi.nii nodif 0 8 > gedit acqparams.txt > > 0 -1 0 0.0608991 > 0 -1 0 0.0608991 > 0 -1 0 0.0608991 > 0 -1 0 0.0608991 > 0 1 0 0.0608991 > 0 1 0 0.0608991 > 0 1 0 0.0608991 > 0 1 0 0.0608991 > topup --imain=nodif.nii --datain=aqu --config=b02b0.cnf --out=my_topup_results This all assumes that the first 8 volumes are b=0, and also that the first four are acquired with an opposing phase-encode direction compared to the next four. If you run a movie on nodif.nii, what do you see? > > fslmaths nodif -Tmean nodif_mean > bet nodif nodif_brain -m It is better to run bet on data that has been corrected for susceptibility. In order to do so you need to augment your topup call with --iout=my_iout and then do fslmaths my_iout -Tmean nodif_mean bet nodif_mean nodif_brain -m > > indx="" > for ((i=1; i<=68; i+=1)); do indx="$indx 1"; done > echo $indx > index.txt This index file implies that you have the same phase-encode direction for all your 68 images. Therefore it is in direct contradiction to the way you run topup above. Can I suggest you check what the 8 first volumes are, and then write back when you have done that? Best regards Jesper > > eddy --imain=dwi.nii--mask=nodif_brain_mask --acqp=acqparams.txt --index=index.txt --bvecs=bvecs --bvals=bvals --topup=my_topup_results --out=eddy_corrected_data > > Then I fed "eddy_corrected_data" to the following steps in the preprocessing pipeline for DTI data in FSL > > Kindly, I have the following questions: > 1. Are the previous steps correct? > 2. Regarding "acqparams.txt". Is it correct for my DTI data ( acquired as mentioned above) to extract all the non-diffusion volumes then create a text file (8 lines one for every b0). The order of the negative and positive values is as above? Is there any difference in the order of the volumes in this file ( I mean adding negatives first or positives first)? > 3. In the file "acqparams.txt" what are 0 1 0 stands for? are these numbers the same in all types of DWI data > 4. Is it correct to feed the output of topup directly to eddy or I need to run the command "applytopup " following "topup" and then feed the output of "applytopup" to eddy? > > Thank you very much for any advice > > All the best, > Jon