Greetings colleagues,
I would very much appreciate your feedback on the accuracy of the following
design matrix, contrast matrix, and group file. I have been using TBSS to
analyze fractional anisotropy data on drug-using subjects at two time points
(time1 and time2). Some of these subjects relapsed between T1 and T2 and
some did not, in addition I have non-using controls. This results in 3
groups of subject. In a traditional ANOVA, the design I would use is a 2x3
mixed model (random effects for group, fixed effects for time). The design
matrix I came up with, for randomize, is the following (for the sake of
simplicity I will just represent the data with 1 control, 2 replased
subjects, and 3 abstinent subjects).
Group/EV1(control time1)/EV2(control time2)/EV3(relapse time 1)/EV4 (relapse
time2)/EV5 (abstinent time1)/EV6 (abstinent time2)/
EV7(subject1)/EV8(sub2)/EV9(sub3)/EV10(sub4)/EV11(sub5)/EV12(sub6)
1/1/0/0/0/0/0/1/0/0/0/0/0
1/0/1/0/0/0/0/1/0/0/0/0/0
2/0/0/1/0/0/0/0/1/0/0/0/0
2/0/0/0/1/0/0/0/1/0/0/0/0
3/0/0/1/0/0/0/0/0/1/0/0/0
3/0/0/0/1/0/0/0/0/1/0/0/0
4/0/0/0/0/1/0/0/0/0/1/0/0
4/0/0/0/0/0/1/0/0/0/1/0/0
5/0/0/0/0/1/0/0/0/0/0/1/0
5/0/0/0/0/0/1/0/0/0/0/1/0
6/0/0/0/0/1/0/0/0/0/0/0/1
6/0/0/0/0/0/1/0/0/0/0/0/1
(the first two lines represent the 1 control subject, lines 3-6 represent
the 2 relapsed subjects, and lines 7-12 represent the 3 abstinent subjects)
Examples of contrasts:
To compare t1 and t2 of the relapsers (simple t-tests):
0/0/1/-1/0/0/0/0/0/0/0/0
0/0/-1/1/0/0/0/0/0/0/0/0
To compare t1 and t2 of abstinent subjects:
0/0/0/0/1/-1/0/0/0/0/0/0
0/0/0/0/-1/1/0/0/0/0/0/0
Here is where I am very confused, how do you model interactions? is the
following correct:
To model a 2x2 interaction with controls and relapsers:
1/-1/1/-1/0/0/0/0/0/0/0/0
-1/1/-1/1/0/0/0/0/0/0/0/0
To model a 2x3 interaction with all groups:
1/-1/1/-1/1/-1/0/0/0/0/0/0
-1/1/-1/1/-1/1/0/0/0/0/0/0
Finally, When I input these matrices into Glm, I get the following error
messages:
Problem with processing the model. Warning: at least one EV is (close
to) a linear combination of the others. You probably need to alter
your design. (Design matrix is rank deficient - ratio of min:max
eigenvalues in SVD of matrix is 1.21792e-17).
Warning-design matrix uses different groups (for different variances),
but these do not contain "separable" Evs for the different groups (it
is necessary that, for each EV, only one of the groups has non-zero
values).
I hope some can shed light on what I am doing wrong, and if possible tell me
what needs to be done to correct these matrices.
Best regards,
Omar Alhassoon, PhD
University of California, San Diego
|