Hello Las,

We have had a chance to look at your data and here is what we have found:

-We ran recon-all on your dataset with higher b-values (30, 40, 50) and found that although it pushed the pial surface out, it reduced the accuracy of white matter segmentation (see 1353_bvalue40)
-We manually edited the temporal lobe area wm.mgz file to fill add more voxels to problem areas, but found that this did not change the pial surface in any substantial way
-We added control points to your original data and found that this was the most effective method of correction. I have attached the control.dat file that I used, and some screen shots of the effects. This method is not perfect and you may still need to play around with number of points and their placement, but it seems to fix the areas you had in question.

In all images:
blue= wm.mgz
green= pial after control points
red= pial original

Images (cps is control points):
-cps_1353_slice76: you can see from this slice that the control point surfaces include much more than the original pial. There are a few problematic spots in the lower part of the image, but you could try editing the white matter mask in this area, or adjusting the control points.
-cps_1353_slice80- same comments as above
-cps_1353_slice84- you can see here that the surfaces with the control points are very close to the original in most areas, but also include more of the temporal lobe in the spot that was giving you trouble.
-cps_1353_slice88- same as above

I am wondering if the problems with the surfaces could be further resolved if you ran the control points on a fresh data set. The one you sent us had already been through recon-all with your adjusted parameters. Were those necessary for your study? Or could you try to just do some editing with control points and a standard recon-all and see what that looks like?

Sincerely,
Ani Varjabedian

On 07/18/2014 09:58 AM, Bruce Fischl wrote:

Hi Las

it's tough to tell from just those images. Can you upload the entire 
subject directory?

cheers
Bruce


On Fri, 18 Jul 2014, Las Blawimo wrote:

Dear dr. Fischl,
 
I was wondering if you've had any chance to look at our images (image 3 and
4 are attached to a second e-mail with the subject: Anterior Temporal Lobe /
Pial Reconstruction Troubles - part 2)?
 
Sincerely,
Las


2014-07-11 16:06 GMT+02:00 Las Blawimo <lasblawimo@gmail.com>:

      Dear Dr. Fischl,

      Thanks for your quick reply.

      Attached you will find some images (image 1&2) showing the
      issues we're having (I'll send a second e-mail including image
      3&4).

      The images are as follows:

      1.) sagittal_1353_norm2+3T_vs_default -> the v5.3 suggested
      norm2 flags (from the release notes) and the -3T flag vs. the
      default v5.3 pipeline. Blue is the norm2/3T flag pipeline,
      yellow is the default.

      2.) sagittal_1353_wusl_vs_default->v5.3 using the -washu_mprage
      flag, obvious improvement in temporal lobe pial surface. Blue is
      -washu_mprage flag, yellow is the default v5.3 pipeline

      3.) axial_1353_norm2+3T_white -> The white surface, which
      appears OK using v5.3 with the suggested norm2 parameters and
      the -3T flag

      4.) axial_1353_wusl_white_error -> the white surface, with an
      obvious error in the lh_postcentral area, using the
      -washu_mprage flag.

      Let me know if you have any suggestions on other parameters I
      should try to improve the temporal lobe pial surface, without
      compromising the white matter surface reconstruction.

      Sincerely,

      Las