Dear all,
I'm migrating from SPM99 to SPM5, and I got some strange results when I
compare two groups of patients. I'm trying to compare groups of brain spects
(around 10-15 pts per group). Most of the time I got 0 difference between
groups. Do you have any experience with SPM5 and SPECT?
Here are the parameters that I'm using:
BASICS DESIGN:
Design: two sample t-test
Grand mean scaling: no
Global calculation: mean
Overall grand mean scaling: yes
independence: yes
Variance: unequal
Normalization: proportional
Anova: no
all the other parameters are the default settings. I tried with ANOVA: yes
and I got another pair of columns in the design matrix, but I got very few
differences between patients.
Thanks
Juan Carlos Quintana
|