Hi,
PACE is excellent because it enables prospective motion correction. MCFLIRT does it retrospectively. However, the MOCO images you get from the scanner are both pro- AND retrospectively motion corrected. This commonly involves some spatial smoothing (as defined by the Filter setting). The non-MOCO series is pro- but not retrospectively corrected. Thus, if you want to use MCFLIRT you should run FSL's motion correction on that time series. Alternatively, use the MOCO without MCFLIRT (and, possibly, no smoothing). Note that Stefan Thesen has implemented 3D-motion correction in k-space which is a very clever idea and may be of advantage. But I have to admit that despite theoretical advantages of the interpolation MCFLIRT just performs so well (i.e., at least equal to the MOCOs) that I continue to run it on the propectively motion corrected data.
Hope that helps,
Cheers-
Andreas
-----Ursprüngliche Nachricht-----
Von: Gonzalo Rojas Costa [mailto:[log in to unmask]]
Gesendet: Mo 17.07.2006 17:54
An: [log in to unmask]; Bartsch, Andreas
Cc: Gonzalo Rojas Costa
Betreff: Re: AW: [FSL] Motion correction in Siemens Symphony...
Hi:
Yes... I ask you the question because Siemens MRI equipments have the
motion correction method named PACE (Prospective Acquisition Correction)
implemented in the equipment, and I think that the EPI images that we
download froom the MRI equipment are motion corrected...
Are that method (PACE) better than another motion correction method such
as the one implemented in MCFLIRT ?... Is better to do the motion
correction in the acqusition stage (PACE) than after that stage (MCFLIRT
for example) ?...
Sincerely,
Gonzalo Rojas Costa
Service of Neuroradiology
Institute of Neurosurgery Dr. Asenjo
Santiago
Chile
|