The identical motion plots are a good sign, and show that the same relative
orientations are estimated both with and without doing the manual
reorientation. i.e. The manual realignment makes no difference.
Basically, the estimated motion that is plotted is derived from the .mat files
(or the sform in the headers in SPM5b) after the images have been realigned.
The plots don't show the difference made by realignment. e.g. if you realign
the same data twice, you should (hopefully) obtain (almost) identical plots
both times.
Best regards,
-John
> I am in the process of analysing an fMRI data set. However, I am having
> problems with the amount of movement (greater than 5 degrees rotation and 5
> mm translation) in some of the subjects. In some instances it is just 1
> scan volume where there is a large movement before returning to the
> original position and in other cases there is a large movement and then the
> new position is maintained for the remaineder of the scanning session.
>
> I have tried to seperately move the position of the images where there is a
> large movement, using the 'reorientate' option before realignment, however,
> even though .mat files are generated during 'reorientation' they appear to
> have no effect on the realignment step (I am getting identical motion plots
> before and after I perform any reorientation of the images).
>
> My question is therefore, is there a way to manually realign specific
> images where there is a large movement to bring them back into line with
> the rest of the data set before realignment?
|