FSL: constructing spherical ROIs and the MNI template

, , ,
1

I am starting to look into spherical ROI analysis in FSL. I am getting into through Andy’s Brain Book tutorial. I was wondering though whether it matters which MNI template is used to index for the voxel coordinates.

I accidentally took the voxel coordinates when loading a 1mm template and applied these coordinates to a 2mm where I noticed I had an erroneous ouput (i.e. no mask was extracted). This led me to wonder whether the MNI template used for ROI mask development should be the same as reported in the paper where the MNI coordinates are reported, or whether the MNI template should be the same as used in the preprocessing of your own data set?

Note: I used the MNI152NLin2009cAsym through fmri-prep, which is not a standard space option in FSL. Also the paper I am pulling the MNI coordinates from does not mention which MNI template they used.

2

Solution: You do not have to use a standard space template to create the ROI. You can use your own data as a “template” to ensure that the voxel dimensions of the max and Zstats align when you go to extract data from the ROI mask. You can check the dimensions with fslinfo.