there is not an easy way to do it. Probably the easiest thing to do is to create a separate session for each run. You can them analyze them under the same analysis name, then combine them together with isxconcat-sess, then do a one-sample group mean (--osgm) fixed effects analysis in mri_glmfit.
On 08/13/2017 09:53 PM, Taylor, Johnmark wrote:
Hello,
I am trying to do a GLM analysis for a study in which I scanned each participant twice. Currently, to combine the two scanning sessions I am combining the runs from the two sessions into a single session directory and applying per-session motion correction, essentially treating all the data as though it came from a single session. The issue is that when I do this, the GLM analysis takes prohibitively long (with just 8 subjects, the analysis has already been running for two weeks). To speed things up, I had the idea of doing the runwise GLM (getting the betas and contrast values for each run), then combining the runwise results to get t-values and p-values for the whole session. However, I am not entirely sure how to do this last bit; is there some kind of function that can be used to take all the per-run t-values or effect sizes in a session and combine them to get an aggregate t-value and p-value for the whole session, for a given contrast? I believe FSL follows this analysis strategy (doing runwise analyses then combining them into the whole session), so I was wondering whether FSFAST had a way of doing the same.
Thank you very much,
JohnMark
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer