Eugenio and Freesurfer team,
Many thanks for responding to these questions. My question relates to hippocampal volume overestimation. I believe that a European team recently evaluated the performance of hippocampus segmentation using different automated
segmentation methods including Freesurfer on more than 400 scans from several previous studies, and found that Freesurfer and FSL-FIRST may include tissue that is not part of the hippocampus. Liedlgruger et al, “Variability Issues in Automated Hippocampal
Segmentation”, 2017, also reported that Freesurfer overestimated volumes vs. other methods tested. I wondered if version 6.0 addresses the overestimation issue? It appeared that the FS 6.0 release web page figure (which showed the hippocampal portion of an
MRI at left, the 5.3 subfield segmentation in the middle, and the 6.0 subfield segmentation at right) showed that a portion of gray tissue superior to hippocampus was included in the 5.3 segmentation but not in the 6.0 version – was that an example of 5.3
overinclusion, addressed in 6.0?
Thank you again for your help and for the team’s continued advances.
Dawn