Hi Leila,

Have you noticed other differences in the same subject using V4 vs V3 (i.e. gray/white bounday differences, thickness thickness difference, etc.) using the same subject? The V4 parcellation should theoretically not be much different from the V3 scheme with the exception of improvement in certain areas such as the anterior cingulate, entorhinal, and the inferior parietal lobule. Where exactly do you find the most amount of difference between the two versions?

Best,

rahul

Hi Leila,

do you see this in other subjects? We also have a utility (mris_divide_parcellation) for automatically dividing each parcellation unit along it's long axis.

Bruce

On Sun, 9 Sep 2007, Leila BH wrote:

...

Hi Bruce, Thanks for your answer, Actually I reprocessed on subject (I have choose one case who doesn't need any editing) and the parcellation result is more accurate in midline but worst in some other part, is it just by chance? That's why I am asking about the variability(in version 3 /and between V3 and V4). I attached 2 tif files for more precision. (one is done with vesion3 and one version4) In my work I have to parcellate the cortex on 1000 small ROIs, I use freesurfer to do that, based on the 35 parcellation, I subdivide each one on small rois, so my small ROIs are not really based on gyrus and sulcus, and to validate my method I want to know the 'error' or variability of the original 35-parcellation. Do you suggest me to run one subject N times to see that? Or do you already know the variability of all the Freesurfer pipeline (till the parcellation)? thanks

-----Original Message----- From: freesurfer-bounces@nmr.mgh.harvard.edu [mailto:freesurfer-bounces@nmr.mgh.harvard.edu] On Behalf Of Bruce Fischl

Sent: samedi, 8. septembre 2007 03:24 To: Leila BH Cc: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] variability

Hi Leila,

Xiao Han's NeuroImage paper discusses thickness variability, and we've looked at parcellation variability in-house, but I don't think we've published on it. The V4 parcellation is should be a bit more accurate at the midline, but not substantially different.

cheers, Bruce On Wed, 5 Sep 2007, Leila BH wrote:

...

Hello,

Is there any paper talking about the variability of the used methods (on reconall )? Especially the parcellation? If I reprocess one subject 10

times

...

with the same Freesurfer version, should I expect the same result for parcellation?

The freesurfer version that I use is stable3

I tried to see if the parcellation of the new Freesurfer version is

similar

...

(visually) to the stable3, and it was very different, is the parcellation method the same in both version?

thanks

---

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