Orientation mismatch DTI2T1

[Bassam Al-Fatly]

Dear Frank,

We have some patients dwi and PET data that we want to use DSI studio for DTI analysis and maybe then some correlational connectometries to clinical or PET scan ROI scores. For that, we thought we would like to do DTI using gqi in native space. The problem is that for our Hammer’s atlas roi to fit on DTI space, we need to coregister to T1 structural and that seems to be a problem since there seems to be an image orientation mismatch. I read in DSI studio website that this is common to problem and can be solved by multiplying by the DICOM orientation matrix with odf_vertices in fib file after recon. The problem now is that we don’t have DICOM rather we have PAR/REC files from Philips. I also tried to extract the orientation matrix from dwi jason file after using mricrogl dcm2nii tool and making third raw of it by X(3,:)=cross(X(1,:),X(2,:)) function in matlab. then I multiplied this matrix by odf_vertices to reorient and saved the workspace to a new fib which I used in DSI later but it failed to fit to T1. We just missed the first step or reconstruction using DICOM table instead of b-table,

any advice? I think it's either not working or we are missing something, MNI reconstruction worked well btw, should we continue then and opt to use MNI reco method instead? in the latter, we should then normalize PET to MNI too, any way doing that together with DTI reco in MNI inside DSI studio?

Thanks in advance and apologies for the lengthy question

Bassam

[Frank Yeh]

Hi Bassam,

If possible, you may send me a sample data and I can figure out a
working solution for you.

Best,
Frank

> --
> You received this message because you are subscribed to the Google Groups "DSI Studio" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to dsi-studio+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/97a9cb41-1113-4897-a1ad-22b3dfdca864n%40googlegroups.com.

[Bassam Al-Fatly]

... thank you Frank, will send you a dropbox shared folder on you official email

[Frank Yeh]

Could you upload the data using the link provided on the DSI Studio website?
https://www.dropbox.com/request/W0hJkAIVC2keWAjqtqvU

> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/1b6bd95b-7614-4c48-b1a5-68f4dc7e0847n%40googlegroups.com.

[Bassam Al-Fatly]

sure, am doing it now :) 

[Frank Yeh]

DSI Studio cannot parse PAR/REC file. Have you tried dcm2niix to
convert PAR/REC to nifti format?

On Tue, Mar 30, 2021 at 3:33 AM Bassam Al-Fatly
<bassam....@gmail.com> wrote:
>

[Bassam Al-Fatly]

I had yes, uploaded the files (nii) analysed with dcm2niix to the bug report dropbox, 
the orientation matrix should be in the json files

[Frank Yeh]

I see. Thank you for the clarification.
Frank

> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/723e0c0b-6c7a-4e03-b58a-c26e6553fea6n%40googlegroups.com.

[Frank Yeh]

I checked your data, and they seem to work on my side:

I use GQI reconstruction at all default settings.

After getting the FIB file, I open it in step t3 and inserts the T1W using [Slices][InsertT1W/T2W]

Then the atlas can be loaded using [Region][Open Region].

If the registration is not good at your end, you may use [Slice][Adjust Registration to fix it].

Frank

[Bassam Al-Fatly]

Wow, great Frank thanks for the update will try it and other patients data and let you know, 

Thanks indeed 

Bassam

[Bassam Al-Fatly]

Hi Frank,
So everything worked well, but for the subjects we don't have par/rec files for, I used the .img .hdr from spm (both dti, T1 and hammers atlas parcellation in native space) using fslmerge to build nii files. I did the whole steps T1,T2,T3 and everything was fine, except for the same issue of T1.nii being having different LR orientation to dti, I used flsswapdim to swap x axis of T1.nii anatomical and hammers.nii images, the trick did the job quite well. But we also wanted to extract info from atlas based automatic tracking from corticostriatal tracts and I donno whether now the dti LR dimension does actually correspond to the HCP fiber atlas in dsi studio, any suggestion?

Thanks in advance
Bassam

PS. read in your shape analysis Neuroimage paper that you also rotate the dti and b table and did quality control for b table rotation! is it pertinent to my question? 

[Frank Yeh]

Hi Bassam,

     I don't have an easy solution here, but I would check the LR orientation of each NIFTI file and make sure that they are correct.

Thanks in advance
Bassam

PS. read in your shape analysis Neuroimage paper that you also rotate the dti and b table and did quality control for b table rotation! is it pertinent to my question?

     Not really. It handles different problem with the b-table.

Best regards,

Frank

 

 

On Tuesday, 30 March 2021 at 21:41:42 UTC+2 Bassam Al-Fatly wrote:

Wow, great Frank thanks for the update will try it and other patients data and let you know, 

Thanks indeed 

Bassam

On Tuesday, 30 March 2021 at 21:39:43 UTC+2 Frank Yeh wrote:

I checked your data, and they seem to work on my side:

I use GQI reconstruction at all default settings.

After getting the FIB file, I open it in step t3 and inserts the T1W using [Slices][InsertT1W/T2W]

Then the atlas can be loaded using [Region][Open Region].

If the registration is not good at your end, you may use [Slice][Adjust Registration to fix it].

Frank

--

You received this message because you are subscribed to the Google Groups "DSI Studio" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dsi-studio+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/b62ae207-d8a6-4da3-a643-19bc35b8b018n%40googlegroups.com.