Dear Sharon
> I'm trying to calculate percentage signal intensity change for a given
> voxel using the "plot" function in SPM96. Could someone possibly tell
> me the easiest way to obtain the values of the maximum and minimum
> plateaus of the boxcar function that SPM assigns to the data?
These are simply the plateau values of the fitted effects in the vector
Y in working memory.
> If global
> normalisation has been used are these values absolute values that can
> directly be used to calculate percent signal intensity change or is
> another step needed?
If global normalization has been implemented to a grand mean of 100
then the difference between the values above is already the percentage
change in terms of global signal.
> Can the resultant percent change values be
> directly compared between independant data sets that have, as closely
> as possible, received identical processing in SPM96?
Yes they can (obviously an inference about this comparison would
require both data sets to be included in the same statistical model).
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|