I'm revisting the problem of large motion in fMRI in order to try to accurately process data collected
from clinical subjects. Some of the movements are more than a voxel, so they are perhaps beyond
the range of the usual motion regressors (Friston, et al 1996) to correct. Grootoonk et al (2000)
proposed a large motion correction method based on adjustment for interpolation errors. This
method was enhanced with Bayesian regularization and implemented in SPM99 , along with
descriptions to be careful when using it. However, it appears that SPM2 implemented new
interpolation methods and unwarping, and I don't see any references to the Grootoonk paper there.
My questions are (1) is the interpolation adjustment proposed by Grootoonk in SPM2?, (2) does the
community think it is good approach for large motion errors?
Paul
Paul Mazaika
Center for Interdisciplinary Brain Science Research
Stanford University School of Medicine
|