Hi Gionzalo, hi Joe-
Joe is certainly right and Jesper has recently convinced me that for things like unwarping motion by B0 interactions, PACE can mess things up.
However, in my experience PACE can "rescue" some of your data in the sense that it can reduce the apparent motion contamination in the data. I also don't find it to hard to get the temporary file with the motion correction parameters PACE is writing out.
So it is not perfect but >can< be useful...
Cheers-
Andreas
________________________________________
Von: FSL - FMRIB's Software Library [[log in to unmask]] im Auftrag von Joseph Devlin [[log in to unmask]]
Gesendet: Freitag, 20. November 2009 16:06
An: [log in to unmask]
Betreff: Re: [FSL] prospective motion correction and mcflirt...
Hi Gonzalo,
This is just my opinion, but I much prefer doing 'motion correction'
post-hoc with FLIRT than using prospective correction. Neither method
actually corrects for motion, though. Instead FLIRT simply realigns
the data whereas PACE reconfigures the gradients in order to
compensate for between scan motion. Both leave 1st, 2nd and hgher
order effects in the data (because these effects are quite complex and
true corrections are extremely difficult to do), but the difference is
that FLIRT gives you the estimated movement parameters whereas PACE
hides these. It is possible to retrieve them but not particularly easy.
Anyway, knowing the amount and type of motion makes a huge
difference. It allows you to make an informed decision regarding the
quality of the data rather than hiding the noise in it. Also, you can
then use the movement information in either the preprocessing (a la
Jesper's realign-and-unwarp or via filtering from something like ICA)
or in the GLM. I prefer these options over prospective correction but
I'd be curious to hear other's opinions.
Cheers
Joe
Joseph Devlin, Ph. D.
Cognitive, Perceptual & Brain Sciences, UCL
Gower Street
London, WC1E 6BT
email: [log in to unmask]
|