If you're willing to use the default resampling option (jac), then the
output from 'eddy' will have one output frame for each input frame, and
you wouldn't need to worry about what an appropriate "average" bvec would
be if you are prioritizing the importance of using rotated bvecs. Of
course, the data will be twice the size, which will have consequences for
the run-times of subsequent processing.
Michael Harms, Ph.D.
Conte Center for the Neuroscience of Mental Disorders
Washington University School of Medicine
Department of Psychiatry, Box 8134
660 South Euclid Ave.Tel: 314-747-6173
St. Louis, MO 63110Email: [log in to unmask]
On 10/22/15 6:53 AM, "Jesper Andersson" <[log in to unmask]>
> 1. Even though i estimated the susceptibility field first using the
>TOPUP tool and then pass the result from that into the --topup parameter
>of EDDY, it does not perform the susceptibility distortion correction at
>all. Hence the merging is also going wrong despite using --resamp=lsr and
that is strange. I have not heard of that problem before. Can you please
send me (to my personal email) the --index file and the --acqp files.
> 2. Afetr performing eddy, you have suggested the average of two
>b-vectors as one possible solution. If one diffusion-weighted image (say
>the first in AP scan ) was corrected for 10 degrees of rotation and the
>corresponding diffusion-weighted image (say the first in PA scan ) is
>corrected for 20 degrees. Why would the average of two rotated vectors be
>valid for the merged diffusion-weighted image (say the first in PA scan
>)? in our example, the average rotation is 15 degrees, and its not a
>valid rotation for the merged diffusion-weighted image.
> Would it be more correct if the b-vectors of AP and PA separately,
>and then calculate diffusion parameter maps (eg FA). And then use topup
>to merge both FA_AP and FA_PA maps? Would it be possible to correct FA
>maps using applytopup using distortion field calculated on b0 maps?
If you see a 20 degree rotation you will have big problems with different
susceptibility fields for the two scans, so data like that will have lots
of problems besides rotating bvecs. For more normal rotations (up to a
couple of degrees) it has quite small effects and an average is probably
At the moment you cannot correct FA maps using applytopup since it will
always do a Jacobian modulation of your images. You will have to decide
what you see as more important, rotating bvecs or doing a better job of
The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.