JISCMail - FSL Archives

Thank you for your useful response and help. Also I apologize for "tar", surely I would use your advice. Actually, I have some doubts which I need your help.

1- Is the exciting of these ghosting would be problematic in next processing steps (for example: tensor estimation and tractography) ??

in general ghosting is not a terrible problem for diffusion as long as the ghosting remains constant during the scan, and as long as the subject doesn’t move. Those are quite big ifs though, so it is always better to try to acquire data with as little ghosting as possible.

2- If it would be problematic, so what do you suggest to solve it ?

There isn’t much you can do at this stage. It is something that should be dealt with in the image recon. There are algorithms for image base ghost correction, but even for those you would need complex images.

3- Do I need GPU in case of using "eddy" (with -jacobian) instead of applytopup ? actually I used and it ran without error just the message "eddy is running in GPU mode" appeared, but the image is completely gray! also changing the rang did not work!

I am very surprised by the gray images. Can you please send me one of those, i.e. upload to the same place as before? The message "eddy is running in GPU mode” doesn’t come from eddy. I guess you are running some script that calls eddy, and that message comes from the script.

4- what did you use as input of the --imain in the "applytopup" ? I mean that, "b0_RL+ all dir of dwi_RL , b0_RL+ all dir of dwi_RL" or all dir of dwi_RL & all dir of dwi_RL" ? since b0 volumes where corrected in topup

I just did a quick test with b0_LR,b0_RL. ALso, topup doesn’t correct anything. It calculates a field that can then be used by applytop or eddy to correct data. If you run eddy you will not need to run applytopup.

5- Some volumes of the DWI image contains bvals =5, am I supposed to split them from the each phase encoded DWI image and merge with corresponding b0 then, perform "topup" on the new merged b0 images?

A bval of 5 is for all practical purposes b0. But I recommend using only a single pair of b0 images for topup even if you have multiple pairs.

Jesper

Regards,

Mahshid

From: Jesper Andersson <[log in to unmask]>
To: [log in to unmask]
Date: Wed, 20 Feb 2019 13:22:39 +0000
Subject: Re: [FSL] ghosting appears when performing topup command on b0 images

Dear Mashid,

I have a data-set including phase encoding (1b0+80 dir RL & 1b0+80 dir LR). Using default config file b02b0.cnf (included in attachment), an image with ghosting around brain was appeared. I also tried to change the values of the --warpres but the results got worse.

I had a look at your data and ran topup on it. The results look perfectly fine to me, given the input data. The ghosting has not appeared as a consequence of topup. There are two bands (one in slices 31-38 and one in slices 62-90) with severe ghosting in your input data. topup only deals with the distortions, and will not do anything about the ghosting.

I also had no problem with applytopup on your data. I think in an earlier version of applytopup the display range was not set correctly, but that would only affect the initial range that fslview (and I assume that fsleyes does the same) displays your data with. Try changing your display range and I think you will see it has worked fine.

For future reference. If you stick all your files in a tar-ball before uploading them you will save time both for me and yourself. If you are not familiar with tar, just google “linux tar” and there will be lots of information.

Jesper

I don't know what is the main problem! I would appreciated if any one could help me with the resulted problem.

Regards,

Mahshid

To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
<b02b0.cnf>

To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1

To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1