Dear SPM experts - I am trying to estimate a level 1 model featuring a
single condition effect modulated by 7+ parametric effects. I am
seeing the following strange behavior: If I have 7 or more parametric
effects for my condition, the last parametric effect entered into the
model always
winds up as a vector of 0s in the filtered/whitened design matrix (as
in SPM.xX.X). This is true regardless of which subject I am estimating
the model for and regardless of the order in which I enter in the
parametric effects. It's also true regardless of whether I enter 7, 8,
or 9 parametric effects. The last effect always winds up as zeros.
I've done/checked all of the following:
1. parameters are de-meaned
2. spm_orth is commented out of spm_get_ons
3. parameters show minimal intercorrelation (max abs corr ~= .20)
4. checked that the input parameters are correct (the parameters are
correctly saved in SPM.Sess.U.P, for instance)
Anyone have any thoughts on what might be going on here? Thanks in advance!
Cheers,
Bob
|