[MINC-users] Regional CBF differences

[Andrew Janke]

> I seek an automated way to determine the quantitative CBF PET changes on a
> region-by-region basis (rather than voxel by voxel).  I have 2 "baseline
> blood pressure" scans and 2 "low-blood-pressure" scans per subject.  I want
> to know how much blood pressure declined/changed in the thalamus, the
> caudate, the cingulate gyrus, and all other "standard" brain regions, like
> those defined by the Talairach atlas.
>
> I'm hoping there's an automated way to calculate these regional CBF
> differences.  I can use mincmath (Baseline1 - LowBP1, Baseline2-LowBP1,
> etc.) to run all 4 possible subtractions for each of the 4 subjects, then
> average these differnece images, again using mincmath.  However, I still
> need some way to analyze the voxels in the difference images (one per
> subject).  I'd like to do the whole brain, and print a table showing these
> average differences, by region, by subject.

OK, so do you have a masked model brain that you plan to use? (be this
the MNI model space -- icbm_152, or some other atlas with defined
areas?)

If so then I would suggest that you used a combination of mincmath (or
minccalc) as you are with mincstats using one of the masks from the
atlas you have chosen.

ie:  (where the caudate is labelled as 10 in an atlas called atlas.mnc

   mincstats -mask atlas.mnc -mask_binvalue 10 subtr-image.mnc

this will print out the mean, median, stdev, sum, etc etc of the local
region that is defined by your mask.


a