Hi Qi Zhu, sorry for the delay (always a good idea to just repost if a few days go by). To answer your question, yes you can do a FFx analysis that way, ie, average the gamma and gammavar across the ROI, then t=gamma/sqrt(gammavar) and use the DOF from the dof.dat file in the mri_glmfit output directory. You'll lose a little bit of power because you will over-estimate gammavar. Simply averaging gammvar over the ROI does not reflect the actual variance of averaged gamma. To get the true gammavar, you'd have to go back to the original time series. But this should work pretty well. doug

Qi Zhu wrote:

Hi,

Sorry for posting again. I just want to get the average ces and cesvar values across subjects (fixed effect) for a specific ROI, and then run some statistics on them. Can I just average the gamma and gammavar that in the osgm folder after I ran the voxel-based ffx group analysis in my group data defined ROI to get those values? And how to average the gammavar values in a ROI? Just mean(gammavar), or sqrt(mean(gammavar^2))?

Thanks very much for the answers.

regards,

Qi Zhu Laboratorium voor Neuro- en Psychofysiologie K.U.Leuven Medical School _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer