My nii without PEinformation

Dear MRtriX Team,
Recently, I downloaded a batch of DTI data in NII format on the Internet. I am using mrcovert to convert the data into MIF format for preprocessing and exporting ‘PE table’. But the error is that phase encoding information is not found. So I used mrinfo to look at my batch of data, but I didn’t find the PE information. How should I solve this problem, or what options should I use, or use a JSON file? thanks a lot.

In the command dwipreproc, I used “- rpenone” and “- pedirAP”. But the error is:

Command:  eddy_cuda9.1 --imain=eddy_in.nii --mask=eddy_mask.nii --acqp=eddy_config.txt --index=eddy_indices.txt --bvecs=bvecs --bvals=bvals --slm=linear --data_is_shelled --out=dwi_post_eddy
dwipreproc: [WARNING] Command failed: eddy_cuda9.1 --imain=eddy_in.nii --mask=eddy_mask.nii --acqp=eddy_config.txt --index=eddy_indices.txt --bvecs=bvecs --bvals=bvals --slm=linear --data_is_shelled --out=dwi_post_eddy
dwipreproc: [WARNING] CUDA version of eddy appears to have failed; trying OpenMP version
Command:  eddy_openmp --imain=eddy_in.nii --mask=eddy_mask.nii --acqp=eddy_config.txt --index=eddy_indices.txt --bvecs=bvecs --bvals=bvals --slm=linear --data_is_shelled --out=dwi_post_eddy
dwipreproc: [SYSTEM FATAL CODE: SIGINT (2)] Program manually interrupted by terminal
dwipreproc: Changing back to original directory (/home/lt292794642/scripts-Jean324)
dwipreproc: Deleting temporary directory /home/lt292794642/scripts-Jean324/dwipreproc-tmp-KGB9WW/

create dwi mask

dwi2mask: [ERROR] failed to open key/value file "/home/lt292794642/adni/connectivity/predwi_denoised_preproc.mif": No such file or directory
dwi2mask: [ERROR] error opening image "/home/lt292794642/adni/connectivity/predwi_denoised_preproc.mif" and so on.

whats the problem?

In addition, I want to know how I can know whether my data is AP direction or LR direction or what direction
thank u very much

Hello @famerfox ,

This is very interesting question. You may find phase encoding directional in the json file. Cheers.

After I downloaded this batch of data, I didn’t find the JSON file, so I was very confused about how to deal with it. Now I use the options “- rpedir” and “- rpenone” instead, but I don’t know whether to use AP or LR after “rpedir”. Thank you for your reply.

Hello @famerfox ,

Would you like to post the weblink from where you downloaded the data?


The error message here indicates that you’ve interrupted the program, presumably by pressing Ctrl+C – there’s no indication of any other error. Are you sure there’s actually a problem?

The previous warning:

tells you that it wasn’t able to run the GPU-accelerated version of eddy, and is falling back to the CPU-based version. This version takes a lot longer to run (hours), maybe you thought the program had stalled…? The chances are it was simply running but taking a long time over it…

As a final note, the options should be: “-rpe_none” (with an underscore) and “pe_dir AP” (with a space) – but I would expect that’s not the issue here since the script wouldn’t have run with these options…

Sorry, I didn’t write it completely. In fact, I wrote “underline” in the script. It’s really because I interrupted the operation ahead of time. I just don’t know the PE direction of my data, which option should I choose, AP or LR or Si. Thank you very much for answering my novice question.
My data comes from: OSF | Calgary Preschool MRI Dataset


If the information isn’t provided with the dataset, you can usually tell by looking at the data. There are examples of what this looks like in for example in figure 1 of this ISMRM abstract, amongst many other sources. But it’s very unusual for the PE direction to be anything other than AP or PA. Otherwise, the options you’ve selected sound correct for that dataset.

Thanks for your reply. My current problem has been solved. My English level is not very good. I may need to ask you more questions in the future. Thank you again.