Dear all,
I'm working on a complex design (possibly too complex). I have a group of 100 patients with a disease A. This disease contains aspects of diseases B, C and D of which I have 25 patients each, which are "pure" B, C or D". I want to know which of the RS-characteristics of disease A are specific for A (e.g. not accounted for by B, C, D).
I've coded each of the factors A-D as 1 (present) or -1 (absent). This means a patients with disease A would get 1 1 1 1 since he/she has all attributes and a patient B would be coded as -1 1 -1 -1.
additionally, I think I need to correct for any interactions between these factors, so I've added factors multiplying each factor for 2-way, 3-way and 4-way interactions. Lastly, I've added age and gender. All were demeaned. I followed the design described under the "ANOVA: 2-factors 2- & 3-levels (2x3 between-subjects ANOVA))" GLM page.
For contrasts, I'm only interested in the main effect of the factor for disease A. The contrast was thus coded as 1 0 0 0 [..] 0 (etc).
However, I'm hitting several problems. First, my design appears to be rank deficient. Can this be explained by the design or am I missing something. Secondly, when running randomise, I'm getting the “Warning: tfce has detected a large number of integral steps. This operation may require a great deal of time to complete. “. First I thought thus made sense since I'm testing a main effect (from zero), but after two weeks it was still stuck at the second permutation.
For randomise I use the command:
randomise -i input -o output -d design.mat -t design.con -m mask.nii.gz -n 5000 -D -T
working on an SGE computer cluster.
Any ideas how to get my design running? Is the design correct in the first place for the question I'm asking?
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|