Hi Alexander,
With high-resolution data such errors do sound a bit too large. It would help to know the exact command line or GUI options that you used. One thing that might matter if coverage of one of your scans is significantly smaller than the whole brain is setting 6 degrees of freedom (you will want to do this anyway, but it should make less of a difference otherwise).
Happy to have a look at your scans if that doesn’t help!
Cheers,
Eelke
> On 8 Jun 2015, at 11:59, Alexander <[log in to unmask]> wrote:
>
> Hi,
>
> When I try to align T1 and T2 images (FOVs are different) using flirt I get ~1.3mm displacement error (assessed based on AC and PC, scan resolution is 0.5 isotropic).
>
> In this specific case I don't really need to use flirt since T1 and T2 are acquired from the same subject and are taken one after another in the same physical position of the subject in the scanner and I can use this info for alignment (q-matrices give me perfect alignment).
>
> My question is whether ~1.3mm error is normal?
> Do I need to pre-process images, eg invert T2 (or T1) (I have similar error when flirt is done on extracted brains although I think in this case other tissue being identical in the same subject should actually help).
>
> Happy to upload the nifti files.
>
> Best wishes
> Alexander
|