FYI


-------- Original Message --------
Subject: Re: [Freesurfer] Thickness repeatability
Date: Thu, 8 May 2008 12:27:35 -0500
From: Xiao Han <elehanx@gmail.com>
To: Doug Greve <greve@nmr.mgh.harvard.edu>
CC: Bruce Fischl <fischl@nmr.mgh.harvard.edu>
References: <200805081511.40394.martin.kavec@gmail.com> <48232A69.1030408@nmr.mgh.harvard.edu> <48232AEA.9020801@nmr.mgh.harvard.edu>


Hi Doug,

How are you?

I couldn't send email to the list. But regarding that two questions, here are my thoughts:
1. I remember we used about 50 iterations. We also have a formula in the paper (Pg 183, right column) that relates the kernel size to the number of iterations.
50 iterations equals 5.6 mm kernel size by that formula.

2. That 6mm (or 50 iterations) is not totally arbitrary. We studied the effect of thickness variability w.r.t. smoothing kernel size. The result is summarized in Fig. 3 of the paper.
Fig. 3 showed that a kernel of 6mm is effective and sufficient to remove noise. Larger kernel will give better repeatability, but may reduce sensitivity in a thickness study.
The point we want to make is that FreeSurfer gives very good repeatability even with a very small smoothing kernel (6mm). That 6mm kernel is indeed much smaller than what other group used when reporting their thickness results (30mm!).  We  made the statement in the 2nd paragraph in the right column of Pg 185.


Best wishes,

Xiao

On Thu, May 8, 2008 at 11:31 AM, Doug Greve <greve@nmr.mgh.harvard.edu> wrote:

Sorry, that was supposed to be a reply to this email


Damian Jenkins wrote:

We have been following the protocol in Han's paper to test repeatability and the following questions cropped up:
1) how many smoothing iterations were run to achieve a smoothing kernel size of 6mm?;
2) what was the reason for choosing this size?

Thanks,
D



Doug Greve wrote:


Probably about 26. You can figure it out from this command:

mris_fwhm --fwhm 6 --niters-only --s fsaverage --hemi lh

dont know how they chose 6

doug

Martin Kavec wrote:

Hi,

is it possible to obtain LGI values of cortices, similarly as curvature indices in ?h.aparc.stats? Could this value possibly be included in the ?h.aparc.stats files?
Thanks in advance,

Martin

 


--
Douglas N. Greve, Ph.D.
MGH-NMR Center
greve@nmr.mgh.harvard.edu
Phone Number: 617-724-2358 Fax: 617-726-7422

In order to help us help you, please follow the steps in:
surfer.nmr.mgh.harvard.edu/fswiki/BugReporting



_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer


-- 
Douglas N. Greve, Ph.D.
MGH-NMR Center
greve@nmr.mgh.harvard.edu
Phone Number: 617-724-2358 
Fax: 617-726-7422

In order to help us help you, please follow the steps in:
surfer.nmr.mgh.harvard.edu/fswiki/BugReporting