Hello,
I'm doing an analysis of SPECT images. I have 2 pool of patients (Depressed
and Control), and several parametric component.
After reading through the SPM book and discussing around the lab, I still don't
see the main difference between :
- Two-sample t-test with my 2 groups as condition and the parametric
component as covariates
and
- Multiple regression with 2 covariates representing my two pool of
patients, and of course also with the paraøetric component as covariates
If I estimate the parameters with both analysis method desribed up here, I
obtain
different results (see attached figures). Not "REALLY" different, but
sensibly different.
I'm curious to know if this difference comes for a theoretical difference I
missed, or from something else that I also missed...
Thanks in advance for your answers.
Christopher Coello
Center of Diagnostic Imaging
Ahus University Sykehus, Norway
|