Rat brain registration HELP

[Byague]

Dear community,

Im totally a newbie and unexpertise in this world of MRI preprocessing.

Im trying to register a rat brain template into my subject brain, to perform voxel analysis, but Im not able to perform the coregister propperly.

Maybe is a problem of the orientation, resolution, if I should applied the registration with a mask... or I dont know how to use the script properly.

Here is the input I use, and the rat brain atlas with its labels.

Im trying using directly antsRegistrationSyN.sh, without any previous step.

May someone could help me?

Balbino
Rat1.zip

Link to template: https://www.nearlab.xyz/fischer344atlas

[ntustison]

First thing you would need to do is get a brain mask of your image. Do you have that?

[ntustison]

A screenshot, generally, is not useful in diagnosing anything more than superficial issues. You're going to need to upload the input data and the precise commands used so that we can reproduce the issue.

[Byague]

Ok, sorry for the screenshot.

template="Fischer344_nii_v4/Fischer344_template.nii"
template_mask="Fischer344_nii_v4/Fischer344_template_mask.nii"

specimen="AlzheimerTemplate.nii"
specimen_mask="brainmask.nii"

fixed_image="fixed.nii.gz"
moving_image="moving.nii.gz"

ImageMath 3 $fixed_image m $template $template_mask
ImageMath 3 $moving_image m $specimen $specimen_mask

antsRegistrationSyN.sh -f moving.nii.gz -m fixed.nii.gz -o test1 -n 16

antsApplyTransforms -d 3 -i Fischer344_labels_v3.nii -r test1Warped.nii.gz -o finalresult.nii.gz -t test11Warp.nii.gz -t test10GenericAffine.mat -n NearestNeighbor

I reorientate the labels from IPL to LPI in itksnap, but when I shut down and turn on the computer labels change to its original orientation.

Sorry for disturbing with such a simple thing, and thank you so much for your help.

Inputs and outputs: https://we.tl/t-CA9bYPnNrJ

[cookpa]

antsRegistrationSyN.sh -f moving.nii.gz -m fixed.nii.gz -o test1 -n 16

Here you've used "moving.nii.gz" as the fixed image and "fixed.nii.gz" as the moving image.

[cookpa]

Also, your labels are not in the same physical space as your fixed image. This is incompatible with registration, the image labels should have the same header (see CopyImageHeaderInformation)

If you load fixed into ITK-SNAP and then the labels (load as a grayscale image, NOT a segmentation), they should be aligned. But they are not:

[cookpa]

After copying the header info

[Byague]

Thousand of thank you both!! Definitely the headache was not using CopyImageHeaderInfo, after that registration was made great!!

[cookpa]

FYI, the Fischer 344 atlas is also available in templateflow along with a good brain mask

[ntustison]

Thanks @cookpa