Hi John
we follow the conventions of the Center for Morphometric Analysis (CMA) here at MGH. I'm not actually sure what landmarks they use to separate what they call "inferior lateral ventricle" from the main one, but you can check with them.
As for ADNI, we might have that data but I'm not sure we are allowed to distribute it. If you check with them about it and let us know we'll give you what we have.
cheers Bruce
On Tue, 17 Jul 2012, John Drozd wrote:
Hi Bruce,
I apologize for my detailist nature. I found the csv files on the ADNI web site containing the numerical values of volume in mm^3 numbers. Do you mean that there should also be the corresponding label maps on the ADNI web site, particularly of the third and lateral ventricles? I know when I ran recon-all it gave all the label maps of each parcellation of the brain and I could overlay these labels on the grey scale images and render 3D surface models in Slicer. If this is the case, could you confirm this. Then I'll email ADNI to see if they have the label maps corresponding to the csv files that they have in the ADNI Study Data -> MRI Images -> Image Analysis link.
I am particularly interested in where Freesurfer separates the lateral ventricles from the third ventricle at the foramina, or at the connection between the posterior horns and the third ventricle. If you have any insight in this, that would be helpful. I am glad that Freesurfer includes the isolated horn pieces, as you explained. I particularly found that accurate segmentation of the horns was very important for the significance statistics in comparing AD, MCI and NEC groups. The next step for me will be sub-regional ventricle volumes, not only left and right but separate volumes of horns. I am curious how you defined the division of left from right for the ventricles? Was it special landmarks on the atlas?
Thank you, John
Sent from my iPhone 4S
On 2012-07-17, at 2:13 PM, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi John
have you checked with the ADNI folks? I think they have FreeSurfer segmentation on many if not most of their subjects. We place no connectivity constraints on the ventricular segmentations because they get so thin that indeed sometimes 1mm data cannot resolve the connections. So sometimes the temporal or occipital horn are disconnected, and sometimes not.
cheers Bruce
On Tue, 17 Jul 2012, John Drozd wrote:
Hello,
I have successfully developed an automated algorithm combining pixel intensity-based and shape based approaches to segment the lateral ventricles. I have applied my algorithm to a small data set of 1.5 T1 ADNI images consisting of 24 AD (subjects with Alzheimer Disease) 19 MCI (subjects with Mild Cognitive Impairment) and 24 NEC (elderly controls). In my cases, particularly AD cases, I noticed that the segmentation of the lateral ventricles is connected to the third ventricle with foramina (usually 1 on a side or on both sides). Also, sometimes I find a connection with the posterior horn and third ventricle. I have researched CSF flow and from the literature, and all of these are real connections. I am near completion of developing an automated cropping scheme with a colleague to crop the third ventricle at these junctions.
Our implementation is a combination of C++ code that we have developed and modified ITK classes. We have been viewing our segmented label maps in Slicer3 and generated 3D renderings using Slicer3's Model Maker module which is based on Marching Cubes.
I have viewed a Freesurfer segmentation in Slicer3 based on a tutorial on the Slicer web site.
I presented a poster of our results at the AIC and AAIC conference which I am at now in Vancouver, and a few people have asked me to compare my volumes with other methods. They were particularly interested in comparing my volumes with Freesurfer ventricle volumes and cross-sectional and longitudinal data. I was told and found csv files on the ADNI web site of Freesurfer analysis. I plan to correlate my values with Freesurfer's values, but I would be interested if Freesurfer also finds these connections between the third ventricle and lateral ventricles and also if it includes isolated temporal horn pieces when the horns are discontinuous due to the MRI not being able to resolve a very thin portion of the temporal horn.
So, would it be possible, if I gave you a list of my ADNI subjects, for you to provide me the freesurfer label maps that correspond to the volumes in the csv files that are on the ADNI web site please? I could run Freesurfer but I would prefer having the label maps that exactly correspond to the Freesurfer ventricle volumes that are on the ADNI website. These would be from the two versions of Freesurfer noted on the ADNI web site in the ADNI Study Data/MRI image analysis link.
Thank you, John
Sent from my iPhone 4S
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.