search

MASILab / Synb0-DISCO

Distortion correction of diffusion weighted MRI without reverse phase-encoding scans or field-maps
https://my.vanderbilt.edu/masi
57 stars 28 forks source link

b0 to MNI registration failing #18

Closed mrribbits closed 3 years ago

mrribbits commented 3 years ago

Hi! So...while my T1s register to MNI fine, the b0 to MNI registration is failing for all my datasets. The b0 ends up way too low (inferior). Then the synthesized b0 comes out poorly of course (screenshots below).

Any ideas why this is happening and/or how to fix? I just tested on 12 subjects and all had the same problem occur.

DWI scan parameters are fairly standard: 256mm FOV; 128x128 matrix; 2mm iso; 80 slices; TE=102ms; GRAPPA=2x.

Sample dataset with synB0 inputs and my outputs: https://www.dropbox.com/s/qllew4f5eketgwv/03C-01.zip?dl=0

b0:

b0

b0_d_lin_atlas_2_5:

b0_d_lin_atlas_2_5

overlay with MNI:

overlay

b0_u:

b0_u

thanks! M

mrribbits commented 3 years ago

After some digging, I see this is a failure of epi_reg. It can be avoided (in my subjects at least) by allowing the --wmseg option so that I can provide the white matter segmentation of the T1 that is output by FAST (T1_fast_pve_2).