Dear SPM experts:
Our lab is doing voxel-based T2 relaxometry data analysis. We collected
32 spin-echo, echoplanar image sets, with TE incremented by 4 msec in each
consecutive image set (e.g., TE 1=32 msec, TE 2=36 msec,¡K TE
32=156 msec) through 10 axial planes (slice thickness=7 mm with a 3-mm
skip, in-plane resolution=3.125 mm ¡Ñ 3.125 mm, field of view=200 mm).
I'd like to complish the analysis by first spatially normalizing one of
the 32 TE-Step image to the T2 template. I picked middle TE-Step T2 image
with TE=88msec, because it shares most similar contrast with the SPM T2
template.
Unfortunately, the normalized image does not come out right, the brain is
stretched and not looking normal. First, our T2 images do not cover up
the whole brain (10 axial slices in a volume, with slice thickness=7mm and
a 3-mm gap), it left out the top and bottom parts. Furthermore, the T2
template in SPM5 has skull in there and my T2 images do not have visible
skull, so SPM stretch the edge of the brain to match the skull of the T2
template. Later I used the EPI template instead and seem to get better
result, but still look bad.
So I have three quesitons:
1. Will the quality of my data (pls see above) allow me to conduct proper
spatial normalization?
2. How do I deal with the partial brain coverage problem? I searched the
archive and people suggest using a full coverage image to coregister to
the partial coverage brain, and then use the full coverage brain to
conduct normalization and apply the matrix to the partial brain. However,
I won't be able to do this since my T1 images are not in good condition
either.
3. Will the function "template weighting image" in SPM5 help me to achieve
better normalization? How do I create the mask? Is there any way I do
not need to manually draw the VOI?
Thank you in advance!
Yi-Shin
|