Consider a T2w sequence instead of T1w for cord seg and registration

[jcohenadad]

The T1w scan, although it is fast and robust for cord segmentation, is not robust for registration given the large anatomical differences in terms of contrast:

T1w

T2*w

I suggest replacing the T1w sequence by a 0.8mm isotropic SPACE sequence (see param here)

[jcohenadad]

Another alternative would be to acquire an MT-like sequence, which would also show similar contrast than the T2w*, and it would also be faster to acquire (~3min vs. 5min for T2 SPACE)