Hmmm, not sure what the issue with that is. I've been trying to get people to stop using curv, so one thing you can do is to save it in mgh format and then use mris_preproc and mri_glmfit instead of mris_glm.
Here's what you can try:
1. Convert a curv file from you subject into mgh: mri_convert lh.curv lh.curv.mgh 2. Load in the mgh file to matlab lhcurv = MRIread('lh.curv.mgh'); 3. Replace the pixel data with your data: lhcurv.vol = myvector; 4. Write the result back out: MRIwrite(lhcurv,'lh.myvector.mgh'); 5. Do this for each subject 6. Run mris_preproc with --meas myvector --srcfmt mgh 7. Then run mri_glmfit
doug
On Tue, 12 Sep 2006, Marie Schaer wrote:
Hello everybody,
I am using the "write_curv" matlab code posted by Doug Greve on 01/10/2003 to convert a curvature vector back into freesurfer .curv format, using the following syntax in matlab:
write_curv('lh.curv',myvector,facenumber);
and when I try to launch mris_glm --meas curv --hemi lh ..., I get the following reading curvature error:
MRISreadNewCurvature: vals/vertex 3 unsupported (must be 1)
... However, I am sure that my vector has the size "verticesnumber x 1", and I also thought that the line 23 of write_curv "fwrite(fid, 1, 'int32');" will anyway assign only one value per vertex.
Could anyone help me? Many thanks,
Marie Schaer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer