Hello,
I am currently doing preprocessing on some diffusion tensor imaging subjects. I have multiple subjects with a bad field of view unfortunately (They have the top of their cortex cut off). When I applied FSL dtifit (tensor fitting) on these subjects, I noticed that the alignement of V1 (longest eigenvector) with the fractional anisotropy (FA) was often badly misaligned to the point I could not use them for TBSS. These subjects don’t have opposite phase encoding and range from 7 volumes (1 b0, 6 gradients) to 32 volumes (1 b0, 32 gradients). To analyze the alignement, I used the enigma FA-V1 alignement protocole on their github but also described here: DTI Protocols « ENIGMA
I think the cause might be the fact I used synb0-dysco as the epi susceptibility correction method which created a synthesized b0 based on a whole brain template which then probably modified the values of the brains with a bad FOV?
I was wondering if anyone had suggestion to what the cause might be? I am trying to limit the amount of scans I eliminate because I am working with some limited and difficulty obtained data.
Thank you,
A