Dear Mark,
If you are performing a registration (not applying an existing transformation) then a 4D image for either the input or reference will be reduced to a 3D image by selecting only the first 3D volume and ignoring the rest.
If you are applying an existing transformation then it will apply the same transformation to all volumes of a 4D input volume, but if the reference image is 4D then it will again just take the first 3D volume and ignore the rest. It is only the temporal dimension of the input image that matters.
If you want to perform motion correction on a 4D image then you should use mcflirt, not flirt.
All the best,
Mark
On 25 Oct 2013, at 20:50, Mark Elliott <[log in to unmask]> wrote:
> Quick question:
> What does FLIRT do when the reference volume is a 4D data file and the input volume is 3D?
> For example, does it use only the first volume of the 4D data set, or perhaps the mean volume?
>
> Thanks.
|