Hi Emma,
yes - you can do this:
run bet first, edit the *_brain.* images you get, and feed these inputs into siena with the lines to 'extract brain' commented out. However, make sure you keep the *_skull* images from bet since siena is using these for calibration / registration purposes (in fact, I would probably not try to further edit those albeit to be totally consistent you would somehow have to). Watch out for the file names, i.e. make sure that siena is using the right files when doing all this. You may also want to consider talairach space masking if you have not already done so. Different brains / heads may need different strategies for optimal bet results even when recorded on the same scanner & by the same pulse sequence.
Please let me point out that I have done the same thing because I have not been happy with the bet results in all cases of one data set. However, in my experience siena is quite robust and the editing to remove remaining non-brain tissue was - at least in my case - not really worth the effort. But you may find it beneficial.
Best regards-
Andreas
-----Ursprüngliche Nachricht-----
Von: Emma Burton [mailto:[log in to unmask]]
Gesendet: Mi 01.09.2004 12:50
An: [log in to unmask]
Cc:
Betreff: [FSL] BET and siena
Hi
I've just started using fsl and I currently have coronally acquired images
that have been converted from dicom to analyze using mricro (256 124 256).
The BET function appears to include alot of non-brain. I have tried to tune
the parameters and change the image centre as suggested in previous emails.
However I still get non-brain tissue in the extracted image. I want to do
serial atrophy measure using siena and don't know how this will affect the
output measure. I was thinking of manually editing the bet images and then
running the siena script. Is this possible? if so which images from bet do
I need to edit and which lines in the siena script do I need to comment out?
I presume to run the siena script I enter the extracted edited image
filenames as input1 and input2?
I hope someone can help.
Thanks
Emma
|