DTI orientation

AFNI Message Board
1

Version ID = AFNI_26.1.00

Hi experts,

I have some DTI and 2 subjects have different brain orientation: AIL vs RPI.
I noticed that in the DEC maps in the QC folder, where the corpus callosum is all blue, instead of red like all other subjects.
But the DWI data of these 2 look correctly oriented in the GUI.

1st I ran this:
dcm2niix_afni -f ${subj}_RL -o DTI/ *dcm
Then I ran TORTOISE.

But going back to the earliest data after dcm2niix_afni I saw the orientation was AIL.

3dinfo says: -orient AIL (see below).
All other subjects are -orient RPI. All the data were collected in the same way...well i think so!! I mean it was the same sequence at the same scanner, but Siemens had issues with DTI

thanks for your help,

Dataset File:    890022_RL.nii
Identifier Code: AFN_rcJ_vHAeMeNWx8gbW8pmVw  Creation Date: Thu Apr 30 09:25:58 2026
Template Space:  ORIG
Dataset Type:    Echo Planar (-epan)
Byte Order:      LSB_FIRST {assumed} [this CPU native = LSB_FIRST]
Storage Mode:    NIFTI
Storage Space:   45,738,000 (46 million) bytes
Geometry String: "MATRIX(0,0,-2,89,2,4.102060e-10,0,-110,-4.10206e-10,2,0,-108):110,110,90"
Data Axes Tilt:  Plumb
Data Axes Orientation:
  first  (x) = Anterior-to-Posterior
  second (y) = Inferior-to-Superior
  third  (z) = Left-to-Right   [-orient AIL]
R-to-L extent:   -89.000 [R] -to-    89.000 [L] -step-     2.000 mm [ 90 voxels]
A-to-P extent:  -110.000 [A] -to-   108.000 [P] -step-     2.000 mm [110 voxels]
I-to-S extent:  -108.000 [I] -to-   110.000 [S] -step-     2.000 mm [110 voxels]
Number of time steps = 21  Time step = 11.00000s  Origin = 0.00000s
  -- At sub-brick #0 '?' datum type is short
  -- At sub-brick #1 '?' datum type is short
  -- At sub-brick #2 '?' datum type is short
** For info on all 21 sub-bricks, use '3dinfo -verb' **
2

Hi, Patrick-

That would be surprising about having 1 subject with different dataset orientation, but things can certainly happen with acquisitions... If you have the original DICOM files still, that would be the best place to start the investigation.

You will want to make sure that the left-right is correct, which is hard to check visually unless you have a marker.

You will also want to make sure that the gradients are still correct. Actually, if this is all the same sequence protocol, what are your gradient values in this dataset with differing orientation?

In terms of checking the gradients, you can use the gradient flip check command, and see if things look reasonable in any of the cases. See here for a general description, and here for an application of it after running TORTOISE specifically.

--pt

3

Hi Paul,

So looking at the raw data, after dcm2niix_afni, I noticed the PhaseEncodingDirection is wrong.
Other correct subject have an: i, -i, j, -j, but the bad subjects have i, j-, j-, j-
Might relate the Siemens issue that changed the DTI acquisition parameters.

890022_AP.json: "PhaseEncodingDirection": "i",
890022_AP.json: "InPlanePhaseEncodingDirectionDICOM": "ROW",
890022_LR.json: "PhaseEncodingDirection": "j-",
890022_LR.json: "InPlanePhaseEncodingDirectionDICOM": "COL",
890022_PA.json: "PhaseEncodingDirection": "j-",
890022_PA.json: "InPlanePhaseEncodingDirectionDICOM": "COL",
890022_RL.json: "PhaseEncodingDirection": "j-",
890022_RL.json: "InPlanePhaseEncodingDirectionDICOM": "COL",

The bvec are different too than the correct subjects:
I'm not sure how to send them to you or show them.

So what can I do?

-pb