Hi,
We either (i) use mcflirt on the whole timeseries with a multi-modal cost function, or (ii) split the timeseries up into tag and control, run mcflirt on these separately, register an example of each together using flirt, and then recombine everything with the appropriate combination of registration matrices. The latter option requires a little scripting but isn't too difficult.
All the best,
Mark
On 16 Aug 2013, at 16:58, Dorian P. <[log in to unmask]> wrote:
> Dear FSLers,
>
> Anybody has experience with correcting ASL data for motion? How do you deal with different signal tag-control intensities and multiple label delays?
>
>
> Any comment might be helpful.
>
> Thank you.
> Dorian
> TJU
|