This would bring QSIPrep more in line with other preps, and would maybe make it easier to support a res-native option for cases where resolution varies across acquisitions, sessions, etc.
@mattcieslak pointed out that most recon methods require isotropic voxels, so we might need something like res-nativemax and res-nativemin to use the largest or smallest voxel dimension, respectively.
For example, if voxels are 1.7 x 1.7 x 1.718, res-nativemax would produce outputs with 1.718 x 1.718 x 1.718 voxels and res-nativemin would produce outputs with 1.7 x 1.7 x 1.7 voxels.
This would bring QSIPrep more in line with other preps, and would maybe make it easier to support a
res-native
option for cases where resolution varies across acquisitions, sessions, etc.@mattcieslak pointed out that most recon methods require isotropic voxels, so we might need something like
res-nativemax
andres-nativemin
to use the largest or smallest voxel dimension, respectively.For example, if voxels are 1.7 x 1.7 x 1.718,
res-nativemax
would produce outputs with 1.718 x 1.718 x 1.718 voxels andres-nativemin
would produce outputs with 1.7 x 1.7 x 1.7 voxels.