Dear AFNI experts,
I am facing a data set of brain tumor patients (T1, DTI and resting state data).
Do anyone have suggestions or experiences on normalizing those brain images with lesions using AFNI?
Thanks!
-lzy
I have used 3dQwarp with the ‘-emask’ option. Here, ‘e’ means “exclude” – that is, the option provides a mask of voxels NOT to use in the alignment process. There are some details you have to understand to use this properly:
[ol]
[li] You have to learn to draw the exclusion mask over the subject’s image dataset, which isn’t hard but takes a little practice. This is done via the AFNI “Draw Dataset” plugin.
[/li][li] The exclusion mask is in the base (template) space, so you have to register the template to the subject’s dataset, and then apply the inverse of the warp transformation to get the subject’s dataset into the template space.
[/li][li] You probably have to skull-strip and 3dUnifize the subject’s dataset ahead of 3dQwarp-ing (to match the template). Scripts @SSwarper or auto_warp.py can do all of these things for you, including the warping, but aren’t appropriate since they don’t allow use of an exclusion mask. So you have to manually assemble the correct processing stream.
[/li][/ol]
I don’t think this is HARD, but then again, I wrote the software and understand how it works better than anyone, so perhaps I’m biased.