[Mristudio-users] ImgVal Calculation

[Hui Jing Yu]

Hi,

I received the following response from Hangyi before (I believe he agrees
with Jun Yi) that the statistical calculation is normalized or weighted by
fiber count.

 

-----Original Message-----
From: Hangyi Jiang [mailto:hjiang at jhmi.edu] 
Sent: Monday, November 24, 2008 1:20 PM
To: hjyu at ic.sunysb.edu
Subject: Re: [Mristudio-users] ImgVal Calculation

 

hi, 

 

this kind of statistics in DTI Studio is based the fibers, not the pixels.
for example, if there are three fibers passing through a voxel, the image
intensity of this voxel will be counted three times when we calcualte the
average image intensity along the fibers.   that is why we named them as
"Statistics ALONG THE FIBERS", not "..along the VOXELS that have fibers".

 

regards,

 

Hangyi

 

 

From: mristudio-users-bounces at mristudio.org
[mailto:mristudio-users-bounces at mristudio.org] On Behalf Of susumu
Sent: Wednesday, December 03, 2008 5:07 PM
To: 'DTI Studio, ROI Editor, Landmarker Questions/Support'
Subject: Re: [Mristudio-users] ImgVal Calculation

 

Good question. No matter how many reconstructed streamlines go through a
pixel (1 or 10), the pixel is "1". It is a binary operation. It would be
interesting to try normalized-line-number-weighted map, but DtiStudio
doesn't do it.

 

  _____  

From: mristudio-users-bounces at mristudio.org
[mailto:mristudio-users-bounces at mristudio.org] On Behalf Of Jun Yi Wang
Sent: Wednesday, December 03, 2008 5:01 PM
To: DTI Studio, ROI Editor, Landmarker Questions/Support
Subject: Re: [Mristudio-users] ImgVal Calculation

 

Hi Dr. Mori,

 

I have a question to your response.  The fiber density (fiber count/voxel)
is usually more than 1, so many line propagations or fibers will go through
the same voxels.  When DTI Studio calculates mean FA for a reconstructed
fiber tract, does it calculate based on all voxels in the fiber tract with
each voxel appear once or based on fibers so that voxels belong to multiple
fibers will be appear multiple times?

Thanks,

Jun Yi

 

  _____  

From: susumu <susumu at mri.jhu.edu>
To: "DTI Studio, ROI Editor, Landmarker Questions/Support"
<mristudio-users at mristudio.org>
Sent: Wednesday, December 3, 2008 3:36:46 PM
Subject: Re: [Mristudio-users] ImgVal Calculation

This is an interesting observation. It should be a simple binary masking.
The pixel with fibers are 1 and other pixels are 0. We multiply this masking
to loaded images and calculate average and STD. Do you think our values are
off? If so, this is definitely something we should look into. Could you send
us (not by MailingList) one fiber file and a co-registered FA map?

 

  _____  

From: mristudio-users-bounces at mristudio.org
[mailto:mristudio-users-bounces at mristudio.org] On Behalf Of Hui Jing Yu
Sent: Friday, November 21, 2008 1:56 PM
To: mristudio-users at mristudio.org
Subject: [Mristudio-users] ImgVal Calculation

 

Dear DTI Studio users,

 

I have a question regarding the ImgVal calculation under statistical profile
for fiber tracking.  I got very different mean FA values (differences
ranging from 0.04 to 0.1)  when I tried these two methods (keeping all
parameters consistent and same version of DTI):

1.       Load previously saved fiber track (binary) into DTI studio, click
on fiber statistics and I have  for Mean of ImgVal and Std. of ImgVal

2.       Using MATLAB, apply the same binary fiber track onto FA image
(analyze format), calculate mean/std values for all the pixels/voxels
belonging to this fiber track

How are these mean ImgVal calculated?  I assume it is just a mean values for
all the pixels/voxels that are contained in the fiber track of interest.

 

Thanks in advance for your help.

 

HJ

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.mristudio.org/pipermail/mristudio-users/attachments/20081203/6e3d0e12/attachment-0001.html