Hi Jürgen,
the transforms found in the base are for each timepoint to the base and vice-versa. You'd need the registrations between the time points, so for example:
tp2_to_tp1.long.base.lta from the tp1.long.base/mri/transform directory (registering tp2 to tp1).
From my test-retest analysis I see that the lta is more accurate than
the -register (internal surface registration) of mris_thickness_diff. But I only tested this on a 15 subjects test population.
Anyway, there is a third way of doing the thickness comparison for longitudinally processed data (and this is what I would recommend, also based on my test with the test-retest data set): As the surfaces in the longitudinal stream are in vertex correspondence across time, you do not need to spatially register at all. Just smooth, subtract, average etc the time points surface measures directly.
So you can e.g. modify Han's method for step 2 similar to this:
mris_thickness_diff -src_type curv \ -out $SUBJECTS_DIR/1001/surf/lh.thickness_diff.mgh \ -nsmooth 50 \ -abs -S 1001 \ -L $SUBJECTS_DIR/1001/surf/lh_thickness_diff.log \ $SUBJECTS_DIR/1001/surf/lh.white \ $SUBJECTS_DIR/1001/surf/lh.thickness \ $SUBJECTS_DIR/1001/surf/lh.white \ <------ ! $SUBJECTS_DIR/1001a/surf/lh.thickness
Note that here I specify the same lh.white surface for both time points. Therefore the registration is skipped and only smoothing etc. applied. Alternatively you can do all this directly with other freesurfer commands (mris_calc ...).
Han's steps 3 and 4 deal with Group analysis and should stay the same independently of cross sectional or longitudinal processing.
Cheers, Martin
On Mon, 2010-02-08 at 18:41 +0100, Jürgen Hänggi wrote:
Hi Martin
Thanks a lot. In the past, I used the -xform /../mri/transforms/.. .lta, but then I saw an email from Doug where he recommended to use -register for longitudinal data. However if I use lta, which transform do I have to use?
Base_to_tp1.lta Base_to_tp2.lta tp1_to_base.lta tp2_to_base-lta
From Base/mri/transform
Or any of the numerous transforms found in the tp1.long or tp2.long folder?
In older versions, I think, there was a kind of combined transform that have to be used.
Do I'am right that when using longitudinal processed data, only step2 of Han's four steps described under
http://surfer.nmr.mgh.harvard.edu/fswiki/ThicknessRepeatibility
have to be run?
Thanks in advance Cheers Jürgen
On [DATE], "Martin Reuter" <[ADDRESS]> wrote:
Hi Jürgen,
I have seen this error before. Not sure exactly what is going on. There could be a bug in Han's code, but I could not find it quickly. I don't recommend the rigid alignment in mris_thickness_diff anyway. I have not seen any evaluation on the accuracy of the registration and do not trust it (it is a rigid surface based alignment using the nearest neighbor approach).
Instead try to specify the lta (transform) for the two corresponding volumes. If you use the longitudinal stream, these transforms should be available (either in the <base>/mri/transform or in the <longitudinal-tp>/mri/transform dir). If you are processing your data cross sectionally: I would recommend to do it longitudinally. If that is not an option use mri_robust_register to register the norm.mgz and use that transform for the mris_thickness_diff.
Cheers, Martin
On Mon, 2010-02-08 at 08:30 +0100, Jürgen Hänggi wrote:
Dear FS experts
when running mris_thickness_diff the following error occurred:
mris_thickness_diff \ ? -src_type curv \ ? -out amblyo_pat01_tp1.long.amblyo_pat01_base1_3/surf/lh.thickness_diff.mgh \ ? -register \ ? -abs -S amblyo_pat01_tp3.long.amblyo_pat01_base1_3 \ ? -L amblyo_pat01_tp1.long.amblyo_pat01_base1_3/surf/lh_thickness_diff.log \ ? amblyo_pat01_tp3.long.amblyo_pat01_base1_3/surf/lh.white \ ? amblyo_pat01_tp3.long.amblyo_pat01_base1_3/surf/lh.thickness \ ? amblyo_pat01_tp1.long.amblyo_pat01_base1_3/surf/lh.white \ ? amblyo_pat01_tp1.long.amblyo_pat01_base1_3/surf/lh.thickness Output differences to file amblyo_pat01_tp1.long.amblyo_pat01_base1_3/surf/lh.thickness_diff.mgh Do a rigid alignment of two surfaces before mapping to each other logging results to amblyo_pat01_tp1.long.amblyo_pat01_base1_3/surf/lh_thickness_diff.log Reading first surface file amblyo_pat01_tp3.long.amblyo_pat01_base1_3/surf/lh.white Surface1 has 137991 vertices Reading in first data file amblyo_pat01_tp3.long.amblyo_pat01_base1_3/surf/lh.thickness Reading second surface file amblyo_pat01_tp1.long.amblyo_pat01_base1_3/surf/lh.white Surface2 has 137991 vertices Reading in second data file amblyo_pat01_tp1.long.amblyo_pat01_base1_3/surf/lh.thickness Register surface 2 to surface 1 (rigid mapping using ICP)
ERROR: FindClosest: MHTfindClosestVertexNo returned annIndex=-1, which is invalid!
What does it mean?
Thanks in advance Regards Jürgen
Jürgen Hänggi, Ph.D. Division Neuropsychology Institute of Psychology University of Zurich Binzmuehlestrasse 14, PO Box 25 8050 Zurich, Switzerland 0041 44 635 73 97 (phone office) 0041 76 445 86 84 (phone mobile) 0041 44 635 74 09 (fax office) BIN 4.D.04 (office room number) j.haenggi[at]psychologie.uzh.ch (email) http://www.psychologie.uzh.ch/neuropsy/ (website) http://www.juergenhaenggi.ch (private website)
This e-mail (and any attachment/s) contains confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorised copying, disclosure or distribution of the material in this e-mail is strictly forbidden.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Jürgen Hänggi, Ph.D. Division Neuropsychology Institute of Psychology University of Zurich Binzmuehlestrasse 14, PO Box 25 8050 Zurich, Switzerland 0041 44 635 73 97 (phone office) 0041 76 445 86 84 (phone mobile) 0041 44 635 74 09 (fax office) BIN 4.D.04 (office room number) j.haenggi[at]psychologie.uzh.ch (email) http://www.psychologie.uzh.ch/neuropsy/ (website) http://www.juergenhaenggi.ch (private website)
This e-mail (and any attachment/s) contains confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorised copying, disclosure or distribution of the material in this e-mail is strictly forbidden.