Hi Jian,
If the striped artifacts are from motion and interleaved slices, it may be possible to repair the
data with the ArtRepair toolbox. It has two ways to catch slice errors...
Try the Noise Filtering option on the raw data. It will automatically find bad slices (often a spin
history effect in interleaved scans), and repair bad slices by interpolation forward and back in time
of just the bad slices. This cuts out bad errors that may spread through slice time correction or
realignment of the volume.
Use the Contrast Movie before and after repairs to see if the repaired image data looks more
reasonable.
And/Or try the Repair and Compare option on the preprocessed data. It will find bad volumes if
spin history significantly changed the global intensity of the volume, and then deweights those
volumes in the estimation process.
Use the Global Quality summary before and after repairs to see if the repaired estimates seem
better.
Software and documentation are at http://cibsr.stanford.edu/tools/ArtRepair/ArtRepair.htm, or
from SPM Extensions. It's compatible with SPM2 and SPM5.
Good luck!
- Paul
Hi Jian
If you use interleaved acquisition, you can get striped patterns like
this in at least 2 ways:
1) Cross talk due to none perfect slice profiles, in this case the
artefact will be there in all volumes
2) Large movements e.g. rotations around the x-axis, or shift in the
z-direction, here the artefacts will be pronounced, but only in a few
volumes around the time point where the movement happened
If you use ascending or descending acquisition you will not se these
patterns, in this case cross-talk will lead to extra smoothing, and I
guess the same is the case with the movement.
As for solutions we have tried to use better slice profiles, and also
inter slice gaps, but with little effect, if you find a solution
please let me know. Oliver Speck and colleagues have solved the
second problem very convincingly.
Best
Torben
Torben E. Lund
Danish Research Centre for MR
Copenhagen University Hospital
Kettegaard Allé 30
2650 Hvidovre
Denmark
email: [log in to unmask]
webpage: http://www.drcmr.dk
[1] O. Speck, J. Hennig, and M. Zaitsev. Prospective real-time slice-
by-slice motion correction for fmri in freely moving subjects. MAGMA,
19(2):55–61, May 2006.
|