definitely use the individuals. Use mri_vol2surf to convert the motion corrected (and otherwise unsmoothed) fmri time courses to the surface. Then use mri_segstats with the --avgwf and --annot options to compute the mean time course over the parcellations.
doug
On Sat, 17 Jul 2010, liang wang wrote:
Hi,
Does anyone know one of studies which uses the parcellations (aparc or aparc.a2009s) Freesurfer provides to extract fMRI time course from each subject's fMRI data. Reading the instructions in Freesurfer Wiki, I know how to do that. However, I am not sure whether I should use the parcellations from the average output (obtained from make_average_subject), or the parcellations based on each subject. I would next compute functional connectivity between these regions. I am appreciated for any suggestions.
Liang