Dear experts,
concerning my question regarding thresholding method of entorhinal and perirhinal labels, I looked at
Augustinack JC et al, Predicting the location of human perirhinal cortex, Brodmann's area 35, from MRI, Neuroimage 2013 Jan 1;64:32-42.
where I found the text:
For each label, the vertices were ordered from most probable least probable), then thresholded so that the surface area of each predicted entorhinal cortex or perirhinal cortex label matched the average surface area of the ex vivo labels.
So, if I understand it correctly, at least for the entorhinal and perirhinal *thresh labels, the threshold has been found individually for each label and hemisphere separately, so that the area of the label in the fsaverage space match average surface area of the corresponding ex-vivo labels? Therefore, no restriction for overlap of perirhinal and entorhinal labels was applied in the thresholding? I found that in the right hemisphere the rh.entorhinal_thresh.label almost completely overlaps with lh.entorhinal_thresh.label.
But, for generating custom aparc+aseg I need non-overlapping labels. I attempted to resolve the issue by using
mris_label2annot --s fsaverage --l fsaverage/label/lh.entorhinal_exvivo.label --l fsaverage/label/lh.perirhinal_exvivo.label --hemi lh --ctab entorhinal_perirhinal.annot.ctab --maxstatwinner --thresh 0.4 --a entorhinal_perirhinal_thresh_0.4
i.e. using --maxstatswinner and --thresh 0.4 which gives non-overlapping labels with shapes approximately similar to the original entorhinal_exvivo.thresh.label and perirhinal_exvivo.thresh.label, but it probably violates the condition of similar area extent.
Could you please comment on if my used method is acceptable, or suggest better method how to achieve non-overlaping entorhinal and perirhinal labels for the purpose of generating custom aparc+aseg from them?
Regards,
Antonin Skoch
Dear experts,
Could you please provide me the information how exactly the *_exvivo.thresh.labels have been thresholded?According to this post, http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg32542.html .thresh labels are thresholded to pick the most likely vertices that give a label that is the right surface area.But how they are thresholded in the boundary of the region where the unthresholded labels do not overlap? There must be additional threshold applied: For example for BA45 there is no further BA label towards frontal pole, but the BA45_exvivo.thresh.label is also restricted in this region wrt BA45_exvivo.label.
I have further question concerning used thresholding method of perirhinal_exvivo.thresh and entorhinal_exvivo.thresh labels. It seems that this method is not applied here since looking at them at the fsaverage surface, they are overlapping. And also *rhinal_exvivo.labels are much more extending to the adjacent regions than *rhinal_exvivo.thresh.labels. What method is applied here?
My main objective to obtain annotation with the vertices properly assigned to perirhinal_exvivo and entorhinal_exvivo to supply this annotation to mri_aparc2aseg to assign ribbon voxels to that parcellation. I think I could use mris_label2annot --maxstatswinner with unthresholded entorhinal and perirhinal labels to assign vertices to most probable label, but I would need to further threshold the labels by correct value to obtain results similar in extent to thresh.labels.
For perirhinal, it seems that the additional threshold is 0.4 (what is rationale for this value?), for entorhinal it seems also around 0.4 but could not find any particular threshold to obtain the exact shape of label corresponding to entorhinal_exvivo.thresh.
Thank you in advance for clarification.
Regards,
Antonin Skoch
I think that is a good method (better than the method I just sent).
On 03/03/2017 07:12 PM, Antonin Skoch wrote:
Dear experts,
concerning my question regarding thresholding method of entorhinal and perirhinal labels, I looked at
Augustinack JC et al, Predicting the location of human perirhinal cortex, Brodmann's area 35, from MRI, Neuroimage 2013 Jan 1;64:32-42.
where I found the text:
For each label, the vertices were ordered from most probable least probable), then thresholded so that the surface area of each predicted entorhinal cortex or perirhinal cortex label matched the average surface area of the ex vivo labels.
So, if I understand it correctly, at least for the entorhinal and perirhinal *thresh labels, the threshold has been found individually for each label and hemisphere separately, so that the area of the label in the fsaverage space match average surface area of the corresponding ex-vivo labels? Therefore, no restriction for overlap of perirhinal and entorhinal labels was applied in the thresholding? I found that in the right hemisphere the rh.entorhinal_thresh.label almost completely overlaps with lh.entorhinal_thresh.label.
But, for generating custom aparc+aseg I need non-overlapping labels. I attempted to resolve the issue by using
mris_label2annot --s fsaverage --l fsaverage/label/lh.entorhinal_exvivo.label --l fsaverage/label/lh.perirhinal_exvivo.label --hemi lh --ctab entorhinal_perirhinal.annot.ctab --maxstatwinner --thresh 0.4 --a entorhinal_perirhinal_thresh_0.4
i.e. using --maxstatswinner and --thresh 0.4 which gives non-overlapping labels with shapes approximately similar to the original entorhinal_exvivo.thresh.label and perirhinal_exvivo.thresh.label, but it probably violates the condition of similar area extent.
Could you please comment on if my used method is acceptable, or suggest better method how to achieve non-overlaping entorhinal and perirhinal labels for the purpose of generating custom aparc+aseg from them?
Regards,
Antonin Skoch
Dear experts,
Could you please provide me the information how exactly the *_exvivo.thresh.labels have been thresholded? According to this post, http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg32542.html .thresh labels are thresholded to pick the most likely vertices that give a label that is the right surface area.But how they are thresholded in the boundary of the region where the unthresholded labels do not overlap? There must be additional threshold applied: For example for BA45 there is no further BA label towards frontal pole, but the BA45_exvivo.thresh.label is also restricted in this region wrt BA45_exvivo.label.
I have further question concerning used thresholding method of perirhinal_exvivo.thresh and entorhinal_exvivo.thresh labels. It seems that this method is not applied here since looking at them at the fsaverage surface, they are overlapping. And also *rhinal_exvivo.labels are much more extending to the adjacent regions than *rhinal_exvivo.thresh.labels. What method is applied here?
My main objective to obtain annotation with the vertices properly assigned to perirhinal_exvivo and entorhinal_exvivo to supply this annotation to mri_aparc2aseg to assign ribbon voxels to that parcellation. I think I could use mris_label2annot --maxstatswinner with unthresholded entorhinal and perirhinal labels to assign vertices to most probable label, but I would need to further threshold the labels by correct value to obtain results similar in extent to thresh.labels.
For perirhinal, it seems that the additional threshold is 0.4 (what is rationale for this value?), for entorhinal it seems also around 0.4 but could not find any particular threshold to obtain the exact shape of label corresponding to entorhinal_exvivo.thresh.
Thank you in advance for clarification.
Regards,
Antonin Skoch
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Antonin Skoch -
Douglas N Greve