>Let me give an explicit example. Take a PET image with a 128 X 128
matrix
and
>63 slices. Let the Global Mean be brain voxels only, the Global Mean
be
50, and
>the grey matter threshold be the default - 80% (0.8).
>
>Is it then correct that
>1) A voxel would have to have a value greater than 40 to be included
in the
>search volume.
> AND a voxel with a value of 25 (50%) of global mean would be excluded
from the
>search volume
>
-This is correct. Note however that a voxel has to have a value larger
than
0.8 times the scan mean in every scan in order to be included. This is
effectively a stricter criteria than saying that the average of this
voxel
should be larger than 0.8 times the average of all scan means. It is
roughly equivalent to vM > 0.8*gM+2*vsd where vM is voxel mean, gM is
the
average global mean and vsd is the voxel standard deviation.
<remainder sniped>
Thank you for your prompt and detailed reply. I am glad to hear that I was not
too far off base.
But your answer raised another question. How does SPM handle missing values
due to pixels falling in and out of the grey matter threshold on different
scans/subjects
Lets say that for pixel Xij (i = subject and j = scan) and a grey matter
threshold of 0.8 we have the following distribution of values across subjects,
with pixel values expressed as fraction of Global Mean
Subj Scan 1 Scan 2
1 1.2 1.4
2 0.75 0.9
3 1.1 1.3
4 0.74 0.78
How does SPM treat this data ? Subj 2, scan 1 falls below the grey matter
threshold and therefore would be excluded from the search volume. On Subject 4,
the pixel falls below the grey matter threshold on both scans. I can think of
several alternatives.
A) Subject 2 and 4 are eliminated from the analysis entirely, i.e. all voxels
are dropped for both of their scans
B) This specific pixel is dropped, leaving 2 subjects for this pixel Xi1, but 3
subjects for pixel Xi2 and all other pixels.
C) This specific pixel is given a value of 0 for Scan 1 and 0.9 for Scan 2 for
Subject 2, and values of 0 and 0 for Subject 4.
D) This specific pixel is given a value of 0.8 (minimum grey matter) for Scan
1 and 0.9 for Scan 2 of Subject 2 and 0.8 and 0.8 for Subject 4
None of these seem exactly right, but B) seems the least problematic except for
having different degrees of freedom for different pixels.
This is not a theoretical issues. I have noted such patterns with FDG scans in
areas such as the medial temporal lobe (amygdala).
Thanks again.
sg
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|