 |
 |
> When I use trilinear or sinc methods, a large chunk of the inferior-most slice > is chopped for many volumes (sometimes the whole slice). Sometimes the > superior-most slice is chopped off instead. Just to be clear, the slices that I > say are chopped still show up as part of the volume--i.e., when viewed, 12 image > frames still show up, but the image in the first frame is cut off or not present > at all. Some have suggested that this may be a masking problem--I have masking > turned off. I tried to add additional inferior and superior slices as suggested > in the MEDx FAQ but this did not help. I think this is unavoidable in the process of correcting for motion, especially when you have a small number of slices. For example, if the head moves upwards during the scan a significant fraction of the slice thickness, then no more data is thereafter collected where the lowest slice used to be. Also, the uppermost slice is now in a position where no previous data had been collected. So, in correcting the motion, if you want to see only the voxels who have data for the entire scan, you are going to have to throw out the voxels that are on the edges of your volume. > Also, is there any objective measure of post-realignment motion (including > rotation)? Center of intensity motion is helpful but I am uncomfortable > ignoring rotation. The SPM alignment function outputs both translation and rotation parameters, listed both in the realignment parameter files created in the image directories and in the graphic output. keith