External Email - Use Caution
Hi Freesurfer Developers,
I am having an issue similar to two posted earlier this year with the visualization of significant clusters from tracula not seeming to match the data. (link to first posting here: https://secure-web.cisco.com/1YIyp3Yo44Yjw3sCRGiQdzImJfJAVEC8lOoyUpu2pWJi9Rv...)
Specifically, when mri_glmfit is first run, the data and initial visualization of the significant voxels along the lh cingulum bundle (dorsal) make it appear that the area with the most significant effect is in the anterior part of the tract (see sig_overlay.png).
When mri_glmfit-sim is run, the text file of the significance values masked by clusters that survive correction also makes it appear that the significant effects would be in the anterior portion of the tract because they are at the top of the text file (see perm.th13.abs.sig.masked.txt), and earlier data, in agreement with information on the tracula page, is given anteriror to posterior along this tract. This matched what I would expect given the data before clusterwise correction.
However, the visualization of the masked significance values seems to show only posterior areas with effects (see sigmasked_overlay.png)
I am wondering if there is something in the mri_glmfit-sim process that woucl be causing this apparent flip in the visualization or if there is something I need to be doing differently.
Please let me know if you need any additional information and thanks in advance for your help!
My code: mri_glmfit --y ${STATDIR}/${TRACT}.avg16_syn_bbr.${MET}.nii.gz --fsgd ${STATDIR}/${STUDY}/inputs/${FSGD}.fsgd doss --C ${STATDIR}/${STUDY}/inputs/${C}.mtx --o ${STATDIR}/${STUDY}/${C}/${TRACT}.${MET}.glmdir --save-eres mri_glmfit-sim --glmdir ${STATDIR}/${STUDY}/${C}/${TRACT}.${MET}.glmdir --cwp 0.05 --perm 1000 1.3 abs
To create visualizations: dmri_trk2trk --in /opt/freesurfer/trctrain/hcp/syn/${TRACT}.display.trk --out ${STATDIR}/${STUDY}/${C}/${ResFol}/${TRACT}.${MET}.maskedoverlay.trk --over ${STATDIR}/${STUDY}/${C}/${TRACT}.${MET}.glmdir/${C}/perm.th13.abs.sig.masked.mgh
To create text file from mri_glmfit-sim makes output: mri_convert --ascii perm.th13.abs.sig.masked.mgh perm.th13.abs.sig.masked.txt
Bryna Goeckner (she/her) Graduate Student Meier Labhttps://secure-web.cisco.com/1L0DdRAsArvZ7gyFkPi0YnO8htwQ03fLvWrtjVgm-rwoJRmJteqw8Dj-iewH4d9zmZvbnFL_5NEgxoJLrGd_NQYfnjFPoLgXMvfK5yej2hwm32-jNm7ZsVzMD8TGhlUepKVqvsgW_zhg97sUkQnOg7IhwG4_pYyB15bh4VO8fXnVZpjJ6LF5z9A75JLJN_nv6kAmlJ3hPmsLa_Y9ScwmTGa8uff94ChjJ6-o_jL_sZV9bhRtu5XqdkQWcQ_pTTcmBwF5qq979w8sRegGIPiYjdOHquZBSDqFiuJxCX3rqFKbUKS0xb4CVwUeC2W7joocNk-Fsh_d8SFPTqYBs6VtRzg/https%3A%2F%2Fwww.mcw.edu%2Fdepartments%2Fneurosurgery%2Fresearch%2Fmeier-laboratory Medical College of Wisconsin Neuroscience Doctoral Program
External Email - Use Caution
Hi Freesurfer Developers,
Reposting this since I haven’t had a response yet -
I am having an issue similar to two posted earlier this year with the visualization of significant clusters from tracula not seeming to match the data. (link to first posting here: https://secure-web.cisco.com/1I4YHjSXDjzJBkl31qnqbYeVzytMt-vQgioQ0jlmsll1FBV...)
Specifically, when following the process detailed on the TRACULA site and mri_glmfit is first run, the data and initial visualization of the significant voxels along the lh cingulum bundle (dorsal) make it appear that the area with the most significant effect is in the anterior part of the tract (see sig_overlay.png).
When mri_glmfit-sim is run, the text file of the significance values masked by clusters that survive correction also makes it appear that the significant effects would be in the anterior portion of the tract because they are at the top of the text file (see perm.th13.abs.sig.masked.txt), and earlier data, in agreement with information on the tracula page, is given anteriror to posterior along this tract. This matched what I would expect given the data before clusterwise correction.
However, the visualization of the masked significance values seems to show only posterior areas with effects (see sigmasked_overlay.png)
I am wondering if there is something in the mri_glmfit-sim process that would be causing this apparent flip in the visualization or maybe it has to do with the visualization process or maybe something else I need to be doing differently.
Please let me know if you need any additional information and thanks in advance for your help!
My code: mri_glmfit --y ${STATDIR}/${TRACT}.avg16_syn_bbr.${MET}.nii.gz --fsgd ${STATDIR}/${STUDY}/inputs/${FSGD}.fsgd doss --C ${STATDIR}/${STUDY}/inputs/${C}.mtx --o ${STATDIR}/${STUDY}/${C}/${TRACT}.${MET}.glmdir --save-eres
mri_glmfit-sim --glmdir ${STATDIR}/${STUDY}/${C}/${TRACT}.${MET}.glmdir --cwp 0.05 --perm 1000 1.3 abs
To create visualizations: dmri_trk2trk --in /opt/freesurfer/trctrain/hcp/syn/${TRACT}.display.trk --out ${STATDIR}/${STUDY}/${C}/${ResFol}/${TRACT}.${MET}.maskedoverlay.trk --over ${STATDIR}/${STUDY}/${C}/${TRACT}.${MET}.glmdir/${C}/perm.th13.abs.sig.masked.mgh
To create text file from mri_glmfit-sim output: mri_convert --ascii perm.th13.abs.sig.masked.mgh perm.th13.abs.sig.masked.txt
Bryna Goeckner (she/her) Graduate Student Meier Labhttps://secure-web.cisco.com/1eDjnPmuyeSrzhfWkOp866rmzsfybJE2zzGXOWTN2ftsxPNG7w-Hdl3NSBqhEN2IMuxZNV1s3WOVDpcmsqw7BuHK51tQibx-P_sEMLJH-WCPjFOOvdD2RF5BT70vRXtx4pgDZAWf925pvhSgudkLNBG1urlIp1Ebv4QvgMmeLfFzTe1NRxDRNLrfa7dKfhndF_B3xikpNBI7n6ASXeQ933LNNBUjtE8_PJmFOjdfpnddwnofk8M6pzb1vTfW4Zb_uwSPPrWzAA3OmPYO8TPEAgHS345eWWia5j9j__mES37BzAmCB2OPjI1p4Bi49ev-D99EzCnFeCd-AJLZJmYEOfA/https%3A%2F%2Fwww.mcw.edu%2Fdepartments%2Fneurosurgery%2Fresearch%2Fmeier-laboratory Medical College of Wisconsin Neuroscience Doctoral Program
freesurfer@nmr.mgh.harvard.edu
-
Goeckner, Bryna