#2 is the safest. I think #1 is probably ok, but you'll want to check that the 12% does not correlate with some factor of interest (eg, all 12% in the control group)
On 9/17/20 8:12 AM, Mikolas, Pavol wrote:
External Email - Use Caution
Dear Freesurfer Developers,
I am analyzing the T1 Data of a multicentric study (N=309, 6 Centers) using Freesurfer. I found out, that most of the scans were performed with the ‘prescan normalize’ option. However, by approx.. 12 % of participants, this option is missing. I also have complete data of all participants WITHOUT prescan normalize. In this analysis, we are interested in the correlation of GM volume in specific ROIs with clinical symptoms. Now we have two options:
1.Use the ‘prescan normalize’ dataset and account for the 12% without this option as a covariate
2.Use the complete dataset without ‘prescan normalize’
It would be very helpful, if you could share your arguments / experience with this, so that we proceed in the best manner.
Thank you, best regards
Pavol
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer