Dear Elkin,
Your age covariate is probably creating trouble here. I would suggest
you specify three different models for all of your questions of interest:
* main effect of G:
within subject, compute the average effect across L and M, ie a contrast
[1 1 1 1 1 1 1 1] (/8) and enter these images, one by subject, in a
two-sample t-test, then use a contrast [1 -1] to test for the main
effect of G.
* main effect of L and GxL interaction:
within subject, compute the differential effect with respect to L, ie a
contrast [1 -1 1 -1 1 -1 1 -1] and enter these images, one by subject,
in a two-sample t-test, then use a contrast [1 1] to test for the main
effect of L and a contrast [1 -1] to test for the GxL interaction.
* main effect of M and GxM interaction:
within subject, compute the average across L for each M level, ie
contrasts [1 1 0 0 0 0 0 0], [0 0 1 1 0 0 0 0], [0 0 0 0 1 1 0 0] and [0
0 0 0 0 0 1 1] then enter these four images per subject in a flexible
factorial design (with factors: subject, G (2 levels) and M (4 levels)
and use contrast:
[1 -1 0 0 1 -1 0 0
0 1 -1 0 0 1 -1 0
0 0 1 -1 0 0 1 -1]
for the main effect of M and contrast:
[1 -1 0 0 -1 1 0 0
0 1 -1 0 0 -1 1 0
0 0 1 -1 0 0 -1 1]
for the GxM interaction.
You can also add age as a covariate for the two first models, possibly
interacting with G.
This should be in alignment with the preprint Martyn sent a link about
yesterday.
Best regards,
Guillaume.
On 06/06/18 15:32, Elkin Luis Garcia wrote:
> Dear experts,
>
> I am performing a mixed anova with group (G, 2 levels; G1=11
> subjects, G2=20 subjects), load (L, 2 levels), and modality (M, 4
> levels) factors, while including age as covariate in the design
> matrix (DM). Additionally, I have included the subject (S) term. I'm
> interested in testing the main effects (G, L, M and S) and their
> interactions (G*L and G*M). My DM contains the following columns: G1
> G2 L1 L2 M1 M2 M3 M4 G1L1 G1L2 G2L1 G2L2 G1M1 G1M2 G1M3 G1M4 G2M1
> G2M2 G2M3 G2M4 AGE S1..S11 S12..S31 (TOTAL=42) As I understood in the
> manual of Gläscher and Gitelman related with Contrast weights, the
> correct contrast to test the main effect of group would be:
>
> 1 -1 0 0 0 0 0 0 .5 .5 -.5 -.5 .25 .25 .25 .25 -.25 -.25 -.25 -.25 0
> ones(1,11)/11 -ones(1,20)/20
>
> However, SPM8/Matlab throws an execution error associated with an
> incorrect contrast specification. Am i making any mistake in the
> specification of the contrast?
>
> Thanks for your answers
--
Guillaume Flandin, PhD
Wellcome Centre for Human Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|