help > RE: 2nd-level analysis
Jun 11, 2015  06:06 PM | Alfonso Nieto-Castanon - Boston University
RE: 2nd-level analysis
Hi Ben,

Regarding (1) yes, you are exactly correct, that contrast will test group x condition interactions in your data (typically for interactions one care about the two-sided results so remember to change the 'one-sided (positive)' option in the GUI to 'two-sided')

Regarding (2) and (3), if you want to do this manually you may select the BETA_Subject*_Condition*_Source*.nii files (these are the first-level fisher transformed connectivity maps for each subject/condition/source) and enter those into second-level anayses in SPM (see the FAQ for some additional info). Also, if you want, you can explore the SPM.mat files generated by CONN for each second-level analysis to perhaps more simply get the appropriate list of files needed (the field SPM.xY contains a list of all the first-level files included into your second-level CONN analyses and the field SPM.xX.X contains the associated second-level design matrix).

Hope this helps
Alfonso
Originally posted by Ben Xu:
Hi,
 
I have two groups (exp and control) of resting state fMRI and each group has two conditions (pre and post intervention). I'd like to look at group x condition interaction at the 2nd-level analysis with one seed ROI. Here are my questions:
 
1. How do I set up the contrast at the 2nd level analysis? Is something like the following correct?
    I selected "See-to_voxel" analysis. For "Between-subjects contrast" [1 -1], "Between-conditions contrast" [-1 1], and "Between-sources contrast" [1] 
 
2. Is it possible to do some of the 2nd-level analyses in SPM8 or SPM12? If it is what kind of the 1st-level analysis files I should use?
 
3. This question is related to Question 2. There are several different types of output files (listed below) after the 1st-level analysis. What are these files? Where can I find the descriptions of these output files?
     - BETA_*_source.nii
     - corr_*_source.nii
     - p_corr_*_source.nii
     - pFDR_corr_*_source.nii
     - Se_Subject*_Condition001.nii
       Se_Subject*_Condition002.nii
 
     Can I use any of these output files for 2nd-level contrasts in SPM?
 
Thanks much for your help.
 
Ben

Threaded View

TitleAuthorDate
Ben Xu Jun 10, 2015
RE: 2nd-level analysis
Alfonso Nieto-Castanon Jun 11, 2015
Ben Xu Jun 12, 2015
Alfonso Nieto-Castanon Jun 12, 2015
Noah Philip Jun 12, 2015
Alfonso Nieto-Castanon Jun 12, 2015
Noah Philip Jun 12, 2015
Ben Xu Jun 15, 2015