Hi FSLers,
Im doing a three group comparison using a one level three factors model
(groups are composed by 15, 37 and 44 subjects respectively). I used the one
factor four levels example as my guide to create my model. So:
EV1 fits condition C (it is the only nonzero EV during condition D). EV2
fits A relative to this, i.e. represents A-C (see below for explanation).
The F-test then tests for any deviation - ie any difference between the
levels, and corresponds exactly to the standard ANOVA test.
I also included demeaned age, sex and handedness as covariates.
When I used randomise executing the same design.con and .mat files from the
Feat ouput I obtained very different results. Why is that?
I read also in the randomise tutorial. There in the 'Repeated measures
ANOVA' it says:
"Following the ANOVA: 1-factor 4-levels (Repeated Measures) example from the
FEAT manual, assume we have 2 subjects with 1 factor at 4 levels. We
therefore have eight input images and we want to test if there is any
difference over the 4 levels of the factor. The design matrix looks like
1 0 1 0 0
1 0 0 1 0
1 0 0 0 1
1 0 0 0 0
0 1 1 0 0
0 1 0 1 0
0 1 0 0 1
0 1 0 0 0
where the first two columns model subject means and the 3rd through 5th
column model the categorical effect (Note the different arrangement of rows
relative to the FEAT example). Three t-contrasts for the categorical effect
0 0 1 0 0
0 0 0 1 0
0 0 0 0 1
are selected together into a single F-contrast
1 1 1
Create a group.dat file for the grouping variable that looks like
1
1
1
1
2
2
2
2
This will ensure that permutations will only occur within subject,
respecting the repeated measures structure of the data."
So if I follow right if one is modelling this type of data the structure of
the design matrix is different from the one used in the Feat example?
Which is:
1 1 0
1 1 0
1 1 0
1 1 0
1 0 1
1 0 1
1 0 1
1 0 1
1 0 0
1 0 0
With contrasts looking like this:
A-D 0 1 0
B-D 0 0 1
Etc...
Thank you very much for your help.
--
Andres Roman-Urrestarazu
|