Hi FSL experts,

I'm having some trouble choosing the correct sign for the unwarp direction when fieldmaps to DTI data. I know that the data was unwarped in the A->P direction. From online reading, I understand there is some trial and error involved with getting the sign (+/-) correct. 

I have tried setting the unwarp direction to both y and y- and have concluded that y- provides better images (see picture in dropbox link).

The problem arises when I get to the TBSS phase of processing. The mean_FA image created by using an unwarp direction of y- for the fieldmap step appears to be somewhat cutoff, which is especially obvious in the coronal view (see picture in dropbox link). Using the unwarp direction of y almost seems to yield a more complete mean_FA, although the image isn't ideal due to the black splotches obvious in the bottom half of the coronal view. 

Can anybody advise here? I believe the y- unwarp direction is providing better DTI images for all subjects, however I'm not sure why the mean_FA image looks truncated or how to fix this. Any advice anybody has would be greatly appreciated!

Thanks a million!

Sheena