Dwipreproc with only PA scan

1

Hi MRtrix experts,
If only PA scan is provided, how can I perform EPI correction without the reverse phase scan? What else can I do to improve T1 alignment later? Can I still use ACT?

Best,
Qi

2

Hi Qi,

If you have the T1, you can use the Synb0-DISCO approach. See here for the publication. I hope this helps.

Best regards,

Manuel

3

thank you Manuel, it is helpful. BTW, do you know how to create b0.nii from dti.nii?
Qi

4

BTW, do you know how to create b0.nii from dti.nii?

Assuming that you intend “b0.nii” to be the mean b=0 image intensity, and you have your gradient table stored externally as an FSL bvec / bval file pair, the command would be:

dwiextract dti.nii -fslgrad bvec bval -bzero - | mrmath - mean b0.nii -axis 3
5

Hi Manuel and Smith,
Thank you so much! I followed the instruction on GitHub and want to verify if I had the right inputs: b0.nii.gz (correct non-diffusion image), T1.nii.gz (raw T1 image) and DTI.nii.gz (raw diffusion weighted image) and acqparams.txt. I have some questions about creating acqparams.txt. If my PhaseEncodingDirection is ‘J’ and the TotalReadoutTime: 0.0575. Then the acqparams.txt would be
0, 1, 0, 0.0575
0, 1, 0, 0.000
Is that correct?

Best,
Qi

6

Hi,

The acqparams.txt looks fine.

One quick comment, I don’t think for this process you need the mean b0, ideally you need to use the first b0 of your acquisition.

Best regards,

Manue

7

Hi Manuel,

As you suggest, so the first b0.nii would be ‘fslroi dti.nii b0 0 1’. Am I correct?

Best,
Qi

8

Hi,

Yes that’s correct (if the first volume of the acquisition is a B0)

Best regards,

Manuel