Thanks Stamatios,

The one 4D diffusion-weighted volume that needs to be reoriented, needs be virtually flipped over the x-axis.

I used this to correct the orieintation:

fslswapdim s11.nii.gz x y -z s11.nii

It was the -z that seemed to correct the orientation. I assume that I now rerun dtifit and check the V1 (lines) to

verify the direction of diffusion which still should be most strongly directional in major tracts

Charlie

On Wed, Jan 29, 2014 at 10:00 PM, Charles Leger <[log in to unmask]> wrote:

Thanks Stamatios,

The one 4D diffusion-weighted volume that needs to be reoriented, needs be virtually flipped over the x-axis.
I used s11.nii.gz x y -z s11.nii

On Mon, Jan 27, 2014 at 5:10 AM, Stamatios Sotiropoulos <[log in to unmask]> wrote:

Hi Charlie

You can use fslswapdim to reorient the data and reverse a certain axis. You then need to swap the sign in the respective coordinate of the bvecs vectors. An easy way to double check things have worked out correctly is to run dtifit, have a look at dti_V1 through fslview and make sure that the fibre orientations (e.g. in the corpus callosum) make sense within all planes.

Cheers
Stam

On 26 Jan 2014, at 18:03, charlie Leger <[log in to unmask]> wrote:

> Preprocessing for a TBSS asymmetries test revealed one participant’s diffusion data has a different orientation (reversed). I assume that, unlike anatomical T1 scans, diffusion-weighted data can’t be corrected at the outset (e.g. dcm2nii –r y file.nii) because the diffusion directional information would not change in tandem with the newly reoriented image, which would disrupt directional information (e.g. FA). Is this correct or can this participants diffusion data set reoriented to be consistent with the other diffusion-weighted data sets.
>
> Any feedback would be greatly appreciated.
>
> Charlie