Group Analysis with Data with different resolutions


I have a dataset where the same protocol was collected on participants in two different scanners. The final resolution of the images from the first scanner was 1.5mm x 1.5 mm x 1.5 mm (after preprocessing- alignment, motion correction registration). The final resolution of the images from the second scanner was 2.5 mm x 2.5mm x 2.5 mm (after preprocessing).

What would I need to do before conducting a group analysis.


Hi Gerome,

Force the voxels to be on the same grid, and blur TO some FWHM level, rather than just adding a blur. That will make the final blur comparable, rather than having the dataset with larger voxels be more blurry. Also, you might consider using scanner as a group-level covariate.

To get the same grid and blur level with, consider options:


Does that seem reasonable?

  • rick

Thanks! I will try it!