Print

Print
Hi Shaoli,

I would remove the whole brain volumes corresponding to the two problematic directions, in order to keep the same number of datapoints across voxels. Removing 2 out of 60 directions should not influence your results dramatically.

You can use fslsplit to get all the 3D volumes and then fslmerge to concatenate only the ones you want to consider. You also need to remove the corresponding entries in the bvecs and bvals files.

Cheers,
Stam  

  

On 31 Oct 2010, at 01:00, shaoli fernando wrote:

Hi,

I have couple of DTI datasets where the participant has moved in 1 or two directions and there is a signal drop-out in 1-2 slices each. Can you suggest how I could exclude these directions/slices and proceed with the analysis. My DTI protocol includes 60 directions and 1 run only.

I am assuming that a fix like above is acceptable.

Can you please help.

Thanks all !

Shaoli