Hello Andreas & list,
I'd like to pick up your question (see below) to the list for which I have not seen an answer so far.
We also read Good paper to copy the VBM model for an aging question in a different population.
On page 5 of her paper it is stated that total grey matter (not tiv) was taken as covariate of no
interest for GM analysis, and total WM for the WM analysis.
I think, but am not abs. sure about this, that agem age2 and age3 were (each) orthogonalised onto the total grey matter regressor.
But maybe sombody else gives a comment on what he/she thinks.
best regards,
Philipp
=====================================
Dear colleagues,
I have some questions about the Good et al. Neuroimage 2001 paper ("A
VBM study of ageing in 465 normal adult brains), especially about some
of the statistical details. I would be grateful if someone could comment
on that.
In their VBM analysis, the authors looked at those brain regions that
were significantly different affected by age compared to the
globalvolume changes. As I understand the statistical model, they used
(principally) an ANCOVA with age as regressor and the total brain volume
as covariate. That is correct? If so, how is it done (technically) in
SPM? The Ancova from "Basic models" does in my opinion not allow such a
model.
And a second question: "age" is not used directly as regressor but
orthogonalized first-, second-, and third order polynomial expansions of
age. Again my question, how is technically implemented in the design matrix.
Best wishes,
Andreas
-------------------------------------------------------
|