> when experimenting with SPM99b, we found that we obtain a map showing
> T-Values as a result. How do these correspond to the Z-Values from
> SPM96 ?
The t values are the direct result of the hypothesis test of a
contrast. Previously these were converted to Z values via something
called the probability integral transform; this amounts to finding a Z
statistic which has the same p-value as the t statistic.
This transformation was applied so that the results on Gaussian random
fields could be applied. Now there are (actually, there have been for
a while) results available for t fields, so the transformation is not
necessary.
-Tom
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|