Hi Doug, I tried the technique you suggested for re-sampling functional volumes from different sessions but from the same subject (and scanner) into the same space, using mc-sess.
To clarify, here we have a situation where the same brain is scanned in the same scanner on more than one occasion (in this particular example, the subjects were taken out of the scanner half way through the lengthy functional recordings for 20 minutes to rest). Obviously the head position and rotation are almost always quite different in session 1 than session 2. It would make building contrasts and combining results across sessions (but within subjects) much more straigthforward if the functional volumes from the two "sessions" could be re-sampled into the same space. I put all my runs (2 sessions, 3 runs each = total of 6 runs) from the same subject and scanner in one bold directory, ran mc-sess, and got confusing results. When I check the alignment with tkregister2, the result looks very bad - the volumes from different sessions are clearly not in the same space (1-3 cm difference). However if I ignore the fact that the runs do not appear to be aligned, and run spmregister-sess just once for all runs (hence there is only one register.dat file) and then check the registration with tkregister-sess, runs from both sessions seem to be similarly and well aligned with the hires anatomicals. Obviously the mis-alignment from mc-sess should be inherited to all following steps, so I must be doing something wrong. I wonder where the mistake is? Am I misusing the tkregister 2 command (see below)? Or is this a bug? *** Here are the specifics: *** Session 1 = runs 004, 005, 006 Session 2 = runs 009, 010, 011 on machine ai source /usr/local/freesurfer/nmr-dev-env setenv SUBJECTS_DIR /space/cognito/5/users/raij/subjects_mri setenv SUBJECT avml19 cd /space/cognito/21/users/raij/avml_fmri_mctogether/avml19_session_BRISI13_ISI1TR_BERT # MOTION CORRECT mc-sess -sf sessid -df sesspar # SPATIAL SMOOTHING AND INORM, OUTSTEM = ifmcsm6 spatialsmooth-sess -sf sessid -df sesspar -fwhm 6 -i fmc -o ifmcsm6 # CHECK ALIGNMENT OF RUN 004 VS. 009 WITH TKREGISTER2 - LOOKS BAD tkregister2 --targ bold/004/fmc.bhdr --mov bold/009/fmc.bhdr --regheader --reg /tmp/junk.reg # RUN SPMREGISTER ONCE FOR ALL RUNS TOGETHER DESPITE THAT ALIGMENT APPEARS BAD spmregister-sess -sf sessid -df sesspar -funcstem fmc # CHECK REGISTRATIONS (ONLY ONE bold/register.dat FILE) - LOOKS OK FOR BOTH, AND VERY SIMILAR FOR RUNS 004 AND 009??? tkregister-sess -sf sessid -df sesspar -funcstem fmc -run 004 & tkregister-sess -sf sessid -df sesspar -funcstem fmc -run 009 & --- Also, the defaults for spmregister-sess and tkregister-sess funcstems are "f". I hope, because my setup does not overwrite the mc-sess output with the output of spatialsmooth-sess, that I am correct in using the "fmc" stem instead of "f"? If I have understood corectly, this has the advantage that I can register motion corrected but not spatially smoothed functional data, which might(?) at least theoretically give better results than registering functionals without motion correction... though the advantage might be too small to be relevant compared to less than perfect registration accuracy. Any thoughts regarding this? I tried the registration with the stem "f" as well, and the problems remained the same, so at least this does not seem to be the cause. Thanks, Tommi > hi doug > > i think i have actually already done what you suggest as my bold directory > consists of all the retinotopy runs (in subdirs 001-010). this is the 'ls' > of > the bold directory: > > 001/ 003/ 005/ 007/ 009/ register.dat same-sess-reg.dat seq.info > 002/ 004/ 006/ 008/ 010/ rtopy/ scripts/ > > i ran the motion correction from the study directory as: >>mc-sess -sf sessid -df sesspar > and it completed successfully. > > -so will that have aligned all the runs together? if so, what can explain > the > fact that the field sign map is not continuous when overlaid on the cortex > and > shows gaps in the blue/yellow colouring? > > thanks! > > jane > > > In message <[EMAIL PROTECTED]> Doug Greve > <[EMAIL PROTECTED]> writes: >> >> You can try putting all the runs into a single bold directory. The >> motion correction will (do it's best to) align them. >> >> doug >> >> >> >> >> >> >> Jane Aspell wrote: >> >> >Hi >> > >> >My question was asked last year by Tommi Raij (see email pasted below) >> but I >> >couldn't find a reply to it in the archives. I am analysing retinotopy >> data >> >that was collected over several sessions but I cannot see how to >> register >> >scans from multiple sessions together, and to the subject's anatomical >> scan. >> >Although Fs-Fast appears to combine the data from multiple sessions >> during the >> >retinotopy analysis it only seems to be possible to use one >> registration file >> >(register.dat) when overlaying the results onto the anatomical. I would >> have >> >thought that registering scans from multiple sessions together would be >> a >> >feature that many people would require. Is there anyway to do it? Tommi >> gives >> >some possible methods in his email below. >> >When I do try and overlay my data onto a surface I see 'gaps' in the >> >blue/yellow colouring of the fieldsign map overlaid onto the cortical >> surface. >> >I wonder if this could be due to the inaccuracy of the registration >> given that >> >I only used the register.dat from the registration of the first >> session. >> >hope that makes sense! >> >i'd be grateful for any help with this! >> > >> >Jane >> > >> > >> > >> > >> >> -- >> Douglas N. Greve, Ph.D. >> MGH-NMR Center >> [EMAIL PROTECTED] >> Phone Number: 617-724-2358 >> Fax: 617-726-7422 >> >> > > _______________________________________________ > Freesurfer mailing list > Freesurfer@nmr.mgh.harvard.edu > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer > Tommi Raij, M.D., Ph.D. MGH/MIT/HMS Athinoula A. Martinos Center for Biomedical Imaging Bldg 149, 13th St Charlestown, MA 02129 U.S.A. [EMAIL PROTECTED] _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
