Translation in Native Atlas Build

[neuropsyched]

When building a native atlas and rendering in 3D I’m getting a large linear translation error that looks like this.

Has anyone encountered this issue before? Any thoughts?

Thanks in advance.

[andreashorn]

Hi there,

two questions:
1. Did you check co-registration and normalization and did both look correct?
2. Did you by chance use FNIRT (FSL)? Because it could be that the inverse FSL routine is currently not working correctly.
-> If not, which normalization method did you use?
How many preoperative files did you use (i.e. was it only the T1 or is there additional data in your project)?

Best, Andy

[neuropsyched]

Hi Andy,

Coregistration and normalization look good. No FSL routines were used. I used ANTs for coregistration and normalization. I have a single preop mri (anat_t1.nii) and a single postop mri (postop_tra.nii).

sorry for the broken link. image here

• This reply was modified 4 years ago by neuropsyched.
• This reply was modified 4 years ago by neuropsyched.

[andreashorn]

This is really weird.

What you could do is delete /patient/atlas/native folder and retry to visualize just to make sure this will be generated again as flawed as before.
If this doesn’t help, if possible, could share the subject with me so I can try to replicate here?

Best, Andy

[neuropsyched]

I’ll give it a try and get back to you.

It almost looks as though the atlas is being registered to the original anat_t1.nii.

Are the anatomy slices in the 3d viewer taken from the normalized preop images (glanat.nii)? I have Preop MRI selected in the drop down of the Anatomy Slice controller. Is it possible that the atlas is being resliced into the original anat_t1 space?

[andreashorn]

Sure, if you are visualizing in native space, it’s all in native space, i.e. the original anat_t1.nii will be visualized, as well.

[andreashorn]

Hi neuropsyched,

sorry, I think this really was a visualization bug – it showed the MNI space backdrop but the atlas in native space. Atlas file should be fine, visualization of the backdrop not.

Please exchange this file:
https://dl.dropboxusercontent.com/u/18255196/ea_assignbackdrop.m

Best, Andy

[akapp]

Hi neuropsyched,

Try re-setting your axis. It may be that part of the image is just cutoff and that’s why it looks weird.

You can get your axis limits by plotting all three slices in a separate figure and typing

V = axis

Go back to your 3D scene viewer and type this into the command window:

resultfig = get(gcf);
% Check that your current figure is correct
if ~isempty(strfind(resultfig.Name,'Electrode-Scene'))
    axis(V) % re-set your axis
end

Hopefully that will fix it. Let me know if you have any problems.

[neuropsyched]

Worked great! Thank you both very much.

[Author]

Posts