Dear Fumiko,
I think such a map does not make much sense, you would not known whether
you look at noise or at a signal. Hard to get published too, I guess +ADs--)
The T values are better for looking at 'real' signal differences, the
exact sizes of T2+ACo- signal don't mean that much in MR images.
When you want it anyway (perhaps to project on a wall at a party +ADs--) ),
you can load your Beta map values (regression coefficients) for your
conditions and the intercept in memory with spm+AF8-vol and spm+AF8-read+AF8-vols,
divide the difference between the Betas of the conditions by the Beta of
your intercept (for all voxels independently), and save the resulting
image again using spm+AF8-write+AF8-vol.
For how to use the above spm+AF8AKg- functions in your own matlab scripts,
type 'help spm+AF8AKgAqACo-' for the function you want to use.
Good luck,
Bas
Op vr, 06-05-2005 te 08:01 -0700, schreef Fumiko Maeda:
+AD4- Hello,
+AD4- I would like to figure out how to get SPM to give me a brain activation
+AD4- difference map in +ACU- BOLD units (rather than, say, a standard contrast
+AD4- map). The analysis is the same as a typical contrast map, but we want the
+AD4- result in signal intensity units, rather than normalized by a variance
+AD4- measure to give something like a t statistic. Is this possible?
+AD4- Any suggestions will be great.
+AD4- Thanks+ACE-
+AD4- Fumiko
+AD4- ----------
+AD4- Stanford Univ.
+AD4- Dept of Psychiatry
+AD4- ----------
|