Hi Ryan, it will depend on the order of the classes in your FSGD file and the order of the +1s and -1s in your contrast. Can you send your FSGD file? If you want to do a post-hoc analysis to see where your interaction is coming from, you can look at csdbase.y.ocn.dat (output from the simulation). From mri_glmfit-sim --help:
csdbase.y.ocn.dat - this is a summary of the input (y) over each cluster. It has a column for each cluster. Each row is a subject. The value is the average of the input (y) in each cluster. This is a simple text file.
You should be able to see which group is thinning and which is thickening.
doug
On 10/24/2012 12:54 AM, Ryan wrote:
Hello everyone, I'm conducting an analysis involving a training group and a control group, where each group has a pre and a post scan. I setup the matrix file, the contrast file, and ran mri_glmfit along with mri_glmfit-sim for the cluster correction. The analysis finished revealing a significant cluster, although I want to confirm that my interpretation of the results is correct. I see a significant cluster which I believe means the training group showed greater thickening compared to the controls, is this correct? Also, how can I tell if clusters represent thickening or thinning in the training group compared to the control group?
When I setup the matrix file, I entered 1's and -1's for the control subjects in the first column, and training subjects in the second column. When I setup the contrast file, I entered a -1 and then a 1, with 0's for the rest of the variables.
Thanks for the help. -W. Ryan McGarry
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer