Hi,
I use randomise to contrast some groups of subjects in a clinical case.
I have 4 groups, where two of them are the subsets of one another: first column = controls, second = patients, third and fourth columns = subsets of patients.
For example:
1 0 0 0
0 1 0 1
0 1 1 0
1 0 0 0
0 1 1 0
0 1 0 0
All goes well until the calculation of the following contrast: -1 0 0 1, while 1 0 0 -1 works well, with the following error:
ERROR: Program failed
An exception has been thrown
Error: tfce requires a positive deltaT input.
Exiting
When I separates the groups, ie by creating one design for the controls/patients and one another for the controls/subsets, everything goes well!
So my question is why? I suspect that the problem came from the fact that variables describing the subsets of patients are strongly correlated with the patient variable (cf. matrix design).
Any idea?
Many thanks.
|