Resample the mask of NMT template by 2*2*2.


Hello, teacher:
When I was studying the macaque fmri, I registered and standardized my macaque image with NMTv2 template, and then I resampled my macaque fmri image by 222. When I resampled NMT's gm,wm,csf and other masks by 222, I found that there were many holes in the sampled mask, and it was not coherent. Is there any way to improve this result?

What command did you use? Did you use something with:
3dresample -dxyz 2 2 2 -rmode NN ...
or something else?

Note that when such a large downsampling occurs (by factor of 4 in all dimensions, from 0.5mm iso to 2mm iso), fine structures might get lost.