Hi David
1. I think the main function of the 3T flag is to align to an Talairach average of 3T data, so probably you should continue to use it.
2. There really is no hard threshold. For the control points you want to use voxels that are entirely wm , no partial voluming and no pathology.
3. You can change some bounds in mri_segment and mris_make_surfaces that will make it more/less aggressive. The help in that command contains some documentation on them. Unfortunately we haven't had the time to document them more thorougly. We'll try to get to it soon
cheers Bruce
On Tue, 27 Nov 2018, David S Lee wrote:
External Email - Use Caution
Hello FreeSurfer Experts! I currently supply inhomogeneity-corrected (via N4 biascorrection) 3T T1w scans through recon-all pipeline. The correction has solved many segmentation errors which reduced the time spent on manual editing. I have several questions related to this:
- I am aware of the "3T" flag that can be supplied to recon-all command. It seems redundant to use
this flag for already-bias-corrected 3T scans (testing from my end also confirms this), but do you recommend any other bias-correction steps?
- What are the minimum and maximum intensity thresholds set for GM and WM during recon-all? In the
official control point tutorial, it suggests that WM should be "close" to intensity of 110, but I wanted to know the exact thresholds to perform more accurate manual editing.
- If these GM and WM intensity parameters can be changed from default, do you recommend doing so?
If there is a documentation I can refer to find answers (I've failed to find one so far), please let me know.
Thank you for your time as always,
-- David S. Lee Research Specialist Center for Healthy Minds University of Wisconsin - Madison (608) 890-1115