Hi Mark,
they are in different coordinate systems, so I don't think you can directly compare them. Maybe Doug has some thoughts, but I don't think it's going to yield reliable results.
sorry :< Bruce
On Fri, 26 Jan 2007, Mark Fenske wrote:
Hi there,
I'm trying to compare regions of activation in functional data collected in two different studies-- one study conducted a few years ago (i.e., before the advent of -autorecon and filled ventricles) and another that was conducted last summer (i.e., all new recons).
The problem is that there are serious distortions in the group-averaged statistical activation maps whenever I use a common surface (e.g., fsaverage) to visualize the results of both studies (details and an example are below). I imagine that this discrepancy is due to the difference in surface geometry around the medial wall of the new vs. old recons.
Is it possible to resample/transform the old surface-based activation maps in some way that accounts for or side-steps the difference in the old vs. new recons, and thereby save us from having to re-run all of the old recons and all of the old fs-fast analyses? Would using mri_surf2vol to go from the old surface to the volume and then mri_vol2surf to go from the volume to fsaverage retain better localization?
Thanks, Mark
Details and example:
Consider this .jpg: http://barlab.mgh.harvard.edu/web_pics/old_recon.jpg
It shows a left medial view with clear regions of activation in medial prefrontal (in white circle) and parahippocampal (in white square) cortices.
It was obtained after running the following commands, :
source /usr/local/freesurfer/nmr-std-env setenv SUBJECTS_DIR /space/odin/12/users/subjects
cd /space/frigg/4/users/mark/average
tksurfer bar_nouchine lh inflated -overlay strvwk_lh_6/osgm/sig.mgh
The next .jpg shows the same contrast resampled onto fsaverage. The distortion in relative localization of activation should be readily apparent (note the empty circle and square): http://barlab.mgh.harvard.edu/web_pics/fsaverage.jpg
I obtained this view using mri_surf2surf to resample and then mri_glmfit to get the activation map:
mri_surf2surf --hemi lh --srcsubject bar_nouchine --sval strvwk_lh_6.mgh --trgsubject fsaverage --tval ./strvwk_lh_6_fsaverage.mgh
mri_glmfit --y strvwk_lh_6_fsaverage.mgh --osgm --glmdir glm_strvwk_lh_6_fsaverage
tksurfer fsaverage lh inflated -overlay glm_strvwk_lh_6_fsaverage/ osgm/sig.mgh
Mark J. Fenske, Ph.D. MGH Martinos Center for Biomedical Imaging - Harvard Medical School Tel: (617) 726-9034 | Fax: 726-7422 | http:// barlab.mgh.harvard.edu/fenske.htm