Sounds like there is something pretty wrong somewhere. As ever, if in doubt try
using the <Check Reg> button to see what the relative positions of pairs of images
appear to be. In this case, compare a pair of images (one that wasn't translated
and one that was) before they are realigned to see if there are any obvious artifacts
that could be responsible for the problem. Then try the same pair after realignment
to see what the .mat files are indicating.
If there are no obvious problems with the data, then you may like to try increasing the
"realignment quality" variable via the <Defaults> button. This will slow things down a little
but you may find things are more robust.
All the best,
-John
> Apologies if this has been dealt with before, but I couldn't find a
> specific reference in the archive.
>
> I have been using a fairly standard fMRI preprocessing procedure, in
> which inspection of the realignment parameter plots has typically
> indicated subject movement over sessions of anywhere between 0-5 mm.
> I have been including these parameters as confounds of no interest in
> the model in order to reduce the likelihood of type I errors.
>
> Recently, however, a colleague and I independently came upon the
> bizarre situation where a single subject appears to have practically
> zero movement across sessions apart from in one session, where he
> apparently moved 50mm in one direction at the start of the session,
> stayed there for 100 or so timepoints, and then moved 50mm back to
> zero again at the end of the session.
>
> Obviously, the subject is unlikely to have undertaken these
> gymnastics without our noticing it. My question, therefore, is what
> might have caused SPM to think there is this much movement, and
> whether there is any way of using the data from this session. Can
> the inclusion of the realignment params in the model correct for this
> degree of apparent "motion"? Is there any other way of convincing
> SPM that in fact this movement does not exist? I have tried
> realigning the session without the first and last few timepoints, and
> there is no difference. I could just leave out this session, but if
> there is an alternative, that would be preferable.
--
Dr John Ashburner.
Functional Imaging Lab., 12 Queen Square, London WC1N 3BG, UK.
tel: +44 (0)20 78337491 or +44 (0)20 78373611 x4381
fax: +44 (0)20 78131420 http://www.fil.ion.ucl.ac.uk/~john
|