Hi Bruce,

Sorry, I did a poor job of describing the problem. 

We are trying to do an analysis of subject specific abnormalities on the surface. Instead of a traditional analysis assuming that patients have homogenous abnormalities (for example cortical thinning in anterior cingulate) relative to a control group, we want to count the number of abnormalities over the entire surface in each subject. On the surface, abnormalities can be defined as average thickness in a small region that is >2 SD away from mean thickness in a control group in that same region. Following that, we can compare the average number of subject specific abnormalities between groups. 

This method has been used to identify white matter abnormalities in patient populations in which heterogenous abnormalities are expected, such as TBI patients. However, it requires an a priori decision about what constitutes an abnormality. In this case, we need to specify it in number of vertices.

We have previously used 128 ul as our defined abnormality size in the volume. Our first thought was to use that as a starting point and translate the volume to surface area or number of vertices.

Is there a reasonable way to approximate how many voxels get translated into surface space? And is there a reasonable way to approximate how much area each data point would have on the surface? 

One approach we are considering is to calculate the average gray matter volume, surface area, and number of vertices from the aparc.stats file. With this information we'd have an approximate relationship between volume, surface area, and # vertices. We are curious if this seems like a reasonable choice, or if you have any thoughts on how to go about this.

Thanks a lot,

Tim

On Fri, Nov 21, 2014 at 11:00 AM, <freesurfer-request@nmr.mgh.harvard.edu> wrote:
Send Freesurfer mailing list submissions to
        freesurfer@nmr.mgh.harvard.edu

To subscribe or unsubscribe via the World Wide Web, visit
        https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
or, via email, send a message with subject or body 'help' to
        freesurfer-request@nmr.mgh.harvard.edu

You can reach the person managing the list at
        freesurfer-owner@nmr.mgh.harvard.edu

When replying, please edit your Subject line so it is more specific
than "Re: Contents of Freesurfer digest..."


Today's Topics:

   1. Re: Repost: estimating lesion size on surface (Bruce Fischl)
   2. Odd thing about mris_convert? Position changed from
      freesurfer format to .stl (ting xu)
   3. prefrontal cortex label (Anna Jonsson)
   4. Re: Odd thing about mris_convert? Position changed from
      freesurfer format to .stl (Bruce Fischl)


----------------------------------------------------------------------

Message: 1
Date: Thu, 20 Nov 2014 16:45:17 -0500 (EST)
From: Bruce Fischl <fischl@nmr.mgh.harvard.edu>
Subject: Re: [Freesurfer] Repost: estimating lesion size on surface
To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu>
Message-ID:
        <alpine.LRH.2.11.1411201644480.6947@gate.nmr.mgh.harvard.edu>
Content-Type: text/plain; charset="utf-8"

Hi Tim

can you elaborate what the issue is? If you have binary lesion maps inthe
volume, can't you just sample those onto the surface and measure surface
area directly? I think I am missing something

Bruce
On Wed, 19 Nov 2014, Timothy
Meier wrote:

> Hello,
> I just wanted to repost this to see if anyone had any thoughts on an
> appropriate 'lesion' size for a pothole analysis performed on the surface.
> We want to compare the average number of small surface lesions (estimated
> from cortical thickness) across single subjects for a control group relative
> to our patient group.
>
> We have done similar analyses in the volume for white matter, so one thought
> was to use that same lesion size. However, that requires estimating a
> surface area from our previously used volume.
>
> Any thoughts on how to approach this, or thoughts on an appropriate lesion
> size to use would be appreciated.
>
> Thanks,
>
> -Tim
>
>