Hi Andrea
1. the control points don't need to go into locations that are misclassified, just ones that the intensity is signifantly different from (and usually less than) 110. They can be correctly segmented as WM.
2. Each control point changes things in a region so they can have a global effects. Not sure how else to answer this - effects can propagate through the volume.
3. Not sure - RMS difference of the output volume?
cheers Bruce
On Tue, 14 Feb 2017, Andrea Gerardo Russo wrote:
Hi FreeSurfer experts, I have some questions (maybe trivial) about Control Points editing.
- Following the guide in your website and exploring your tutorial subject
"cp_before" I understood that a Control Point has to be placed in a region that is definitely white matter but the standard recon-all process has classified as gray matter. However in the section "General Tips" you suggest that a Control Point has not to be placed in a tissue different from the white matter, and in other questions posted on the this mailing list several scientists said that they place Control Points in the white matter voxels with an intensity lower than the 110 to adjust the local intensity. I'm a bit confused. What is the correct strategy? 2) I noticed that a Control Point can adjust the white matter surface in its neighborhood, but it can create unpredictable changes in surfaces in other regions. How is it possible? 3) What kind of index can I use to test the reliability of Control Point editing? In a test-retest protocol, considering a single subject, I noticed that although the Control Point positioning improves the FreeSurfer surfaces there is an increase of the covariance in the cortical thickness measurements between scans after the editing. How can I interpret this results?
Best regards Andrea Gerardo Russo