Hi all,
I have DWI data for >100 subjects. I'm transforming FA maps from (1) native space to T1 space using FLIRT and then from (2) T1 space to MNI152 space using FNIRT (and a premat FLIRT affine-transformation).
My script worked for all subjects except for a single one. This subject's T1 image looks fine, and the orientations/axes (left/right, superior/inferior, posterior/anterior) looks fine. However, after FNIRTing the T1 to MNI152 space, the resulting output has very bad warping. And, I can see that the 'views' seemed to be swapped.
In more detail: when viewing the standard MNI152 in FSLeyes, we see that the left-most window consists of sagittal slices, the middle window consists of coronal slices, and the right-most window consists of axial slices. However, when viewing FNIRT's output, the left-most window consists of coronal slices, the middle window consists of axial slices, and the right-most window consists of sagittal slices (all of which are very badly warped). This exact same thing occurs when just using FLIRT as well. Not only do the windows have the wrong slice-view, but the brain images seem to be rotated as well. For instance, the coronal slices (which are incorrectly presented in the left-most window/the sagittal window) are rotated -90 degrees. Same with the axial slices (incorrectly presented in the middle window/the coronal window). And the sagittal slices (incorrectly presented in the right-most window/the axial window) are flipped -180 degrees.
I'm confused because the axis labels and orientations for this subject's T1 are fine (we use a fish-oil pill in the head coil to indicate the right side).
My script is fine considering that all 100+ other subjects' transformations look perfect. Does anyone have any idea of what is going wrong?
Thank you!
|