Hi,
Sorry for the following onslaught of naive q's....
I've been trying to modify a Matlab script to do a random effects analysis with a single group in SPM 8. During the first level analysis, I have defined additional contrasts. In the current state of my group analysis script, the relevant con*.img files for a given contrast across subjects are loaded, and the script estimates parameters (spm_spm). Now I'd like to see what the same contrasts I have defined on the first level are like on the group level.
So I assume I need to now define contrasts as well and call spm_contrasts; however, now in each of my design matrices (one for each t-stat done during 1st level analysis), the column is only a single mean (and each row = each observation, i.e. respective contrast for each subj). Does this mean my contrast for each 1st level contrast should be something like [1 0] (which doesn't work)? If I had more than one group, would I have to zero-pad (# groups - 1) times? Or do I now need to load the beta img generated from the 2nd level parameter estimation for each 1st level contrast and work with that?
I'm not sure if I understand where/at what point/how I can specify the t-threshold or extent threshold either. Do I have to make an explicit call to spm_VOI after spm_contrasts?
If someone has an example script (not using batch jobs) for second level analysis, that would be really helpful, too...I've already trawled through the web and of course found Rik Henson's examples, the AA modules, etc...but I'm still stuck :(
Thanks in advance!!!
g
|