on the acquisition part, online motion correction might help (i.e. estimating the movement between the second but previous and the previous scan online and adjust the recording of the current volume by that).
Stefan Thesen has written a Siemens EPI with such prospective acquisition correction (PACE) which performs quite well. Jespers tool is nice and you can easily use that as Joseph was pointing out and proceed with FEAT. However, if your data contains very little movement the method may not perform well.
Overall, I want to point out that any of Meredith's approaches may be worth a trial even though Joseph's objections apply.
Von: Joseph Devlin [mailto:[log in to unmask]]
Gesendet: Fr 24.09.2004 23:04
An: [log in to unmask]
Betreff: Re: [FSL] motion questions
Yep, that's the hard part - some populations are just difficult to get good
data on and throwing their data away is not really an option in some cases...
Using motion as covariates of no interest is a good idea except with
stimulus correlated motion. But having seen Jesper's software in action,
I've been extremely impressed. I now use his code with essentially all of
my data and no longer bother with motion covariates. I believe we are
planning to add this code to FSL in the not-too-distant future, but for the
moment, it's not an option in Feat.
Joseph T. Devlin, Ph. D.
FMRIB Centre, Dept. of Clinical Neurology
University of Oxford
John Radcliffe Hospital
Headley Way, Headington
Oxford OX3 9DU
Phone: 01865 222 738
Email: [log in to unmask]