Dear List,
I am running a couple of analyses with single or multiple different regressors of interest and have been observing the beta estimability issue with some subjects. The topic is the same as in message 034661 in the archives. The thing that I particularly don't understand is why SPM8 drops some sessions for some subjects, even when the only modulator I put in is fully independent from everything else. E.g. I have the problem when only using RT as a regressor of interest and the task and motion regressors as of no interest. I am using a direct call to the functions mentioned on the archive message, but I'm not sure this helps, given that the model itself doesn't include the column of interest for the problematic sessions in some subjects.
> SPM.xCon(n) = spm_FcUtil('Set',cname,'T','c',c,SPM.xX.xKXs);
> spm_contrasts(SPM);
The output screen in matlab shows that SPM is adjusting something in those problematic columns and the results look normal. But I don't understand what's going on. Is there a way I can verify this is not giving me incorrect results? Also, what exactly is going on? Why might the same sessions for some subjects tend to drop different regressors?
Thanks for any help!
--
Christian A. Rodriguez
Graduate Student
Decision Neuroscience Lab
Department of Psychology
Stanford University
|