Dear Christian,
also make sure that you don't mistake Kalina's script: Her intention was to
supply an extended mask for stats inclusion. She used grey as well as white
matter estimations for it, combined both and performed additional smoothing.
For example, I have found her script valueable to analyze data for a patient
who suffered from a praecentrally located angioma. Due to the relative
hypointensity in the angioma vessels and their flow void, lots of it and the
immediate surroundings would have been masked out and I did want to keep it.
But her script was not to limit the stats to grey matter. In any case, this
is usually detrimental to the Gaussian random fields. Thus, in spm such
masking should mostly be done once the stats were calculated.
Best regards-
Andreas
----- Original Message -----
From: "Karsten Specht" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Monday, March 17, 2003 6:26 PM
Subject: AW: binary gray matter mask at a p level of 0.4
Dear Christian,
> I want to replicate the work of Lipschutz, Friston,
> Ashburner, Turner &
> Price (2001). Assessing Study-Specific Regional Variations in fMRI
> Signal. NeuroImage, 13, 392-398.
>
> Referring to this I have a question on gray matter masks:
> How can I generate a "binary gray matter template thresholded at a
> probability level of 0.4" (p. 394) from the SPM99 EPI template? As far
> as I know the standard mask in SPM99 has a different
> probability level.
> Kalina Christoff wrote a matlab routine that generates gray matter
> masks.
> See:
> http://www-psych.stanford.edu/~kalina/SPM99/Tools/glm_specmask.html
> I guess a simple modification of her script should do it.
> Any helpful suggestions would be appreciated.
I'm not completely sure, and I didn't checked the paper, but I think, they
masked with the gray matter template, which is in the apriori directory. The
voxel values of this image represents the probability of being gray matter.
To get only voxels with a probability level of 0.4 you can use the ImCalc
function by entering i1>0.4 as formula. The resulting image (written into
the current working directory) will be the binary mask you are looking for
But anyway, you can get such a image also from own data by segmenting them.
The '*_seg1' image is such a gray matter image and you can perform the
ImCalc procedure also on that image.
Good luck,
Karsten
----------------------------------
Karsten Specht
fMRI Section
Department of Neuroradiology
Medical Center Bonn
Spessartstrasse 9
53119 Bonn
Germany
Phone: ++49-(0)228/90 81-178
Fax: ++49-(0)228/90 81-190
E-Mail: [log in to unmask]
http://www.mcbonn.de/Praxis/praxis15/fmri1.htm
http://www.bergen-fmri-group.org
|