I have a question as to how the global values are derived. I ask because I
have noted that the global values that are listed on the first output page in
spm.ps are much smaller than values I obtain on my own by segmenting the brain
from extrabrain voxels, and then taking the mean of all pixels in the brain that
have a value greater than 0.
For example, the SPM global listed for a subject = 486.531 (using proportional
normalization, no grand mean scaling, grey matter threshold = 0.7) whereas I
obtain a value of 673.1 (these are FDG values). Is the global mean given in
the SPM output the mean of all brain voxels or the mean of all brain voxels that
are above the specified gray matter threshold. If the latter is true, then how
can SPM obtain a mean of 486, that is only slightly higher than the cutoff value
of 471 derived from the grey matter threshold setting of 0.7 (0.7*673 = 471) ?
Inspection of the image shows that most of the cerebral voxels have values of
800 -1800 in this scan.
The only thing that makes sense to me that the listed "Global" value is is the
criterion value for the grey matter threshold, i.e., the value that a pixel must
exceed to be considered as grey matter, and not the mean of all pixels in the
brain or even the mean of all grey matter pixels.
A second question is whether the global values listed in page 1 of spm.ps are
the values used for proportional normalization ? If not, then how is the global
value used for proportional normalization of each scan calculated ?
Thanks
sg
====================================
Steven Grant, Ph.D.
Cognitive Neuroscience of Addiction Program
Etiology and Clinical Neurobiology Branch
Div. Clinical and Services Research
National Institute on Drug Abuse
Room 4-4238
6001 Executive Blvd
Bethesda, MD 20892
301 443-4877 (voice) 443-6814 (fax)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|