This has to do with the labeling around the medial wall, which is asymmetric in bert (and probably other subjects). In bert's left side, the medial OFC wraps around the lateral OFC to meet the insula. On the right side, the insula wraps around to reach up to the the MOFC. The wmparc is generated by mapping a given WM voxel to the closest point in cortex, so these cortical labeling errors/asymmetries transfer. I'm not sure there's much that can be done about it other than editing the aparc.
On 03/06/2017 03:27 PM, Duncan Cleveland wrote:
Dear FreeSurfer experts,
I am writing today to ask if you can help my group and I with an issue with the white matter parcellations. While inspecting the output from wmparc.mgz in our dataset, we noticed an unusual parcellation in the medial- and lateral OFC labels in both the left and right hemispheres. Specifically, the white matter labels often extend behind the nucleus accumbens and the putamen (see attached screenshot FS_OFC.tif). Curiously, there appears to be lateralization in the extension; while the right hemisphere often has both the medial and lateral OFC labels extend, the left hemisphere only sees the medial OFC label extend that far back. We have looked through several subjects and their FreeSurfer output, and the effect appears in nearly all the subjects.
Initially, we thought this was an issue with our processing stream or subtle biases in our data collection. However, subjects which come packaged with FreeSurfer also show the same effect. Bert, for example, shows the same extension of the OFC posterior to subcortical structures (see screenshot Bert_FS5_OFC or coordinates 133, 127, 137). Again, we see the differences in hemispheres as well, with the left containing only medial OFC posterior to subcortical structures and the right containing both lateral and medial OFC. We are running version 5.3, but the issue persists into version 6.0 (Bert_FS6_OFC or coordinates 134, 128, 137). In fact, we not only see hemispheric differences, but the left hemisphere medial OFC actually bleeds into the right hemisphere as well. Fsaverage does not have a wmparc, so I was unable to see if the issue persists in that output as well.
I have a number of questions associated with this issue. First, to confirm, is this a real issue of mistaken anatomy? The output in this region simply looks incorrect to us anatomically, as we believe that the OFC should not extend posterior to subcortical structures. Second, is there a method to restrict the OFC to regions anterior to the nucleus accumbens without messing up the rest of the algorithm? And third, why do we see differences in hemispheres? Why does the left hemisphere only have medial OFC white matter posterior to the accumbens when the right hemisphere has both?
Thanks and best, Duncan Cleveland MCTFR University of Minnesota
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer