Hi Charlotte
the aim of the WM intensity normalization is to center the white matter class mean around 110 and to "flatten" it over space, so that the mean is spatially stationary. It is certainly harder with 1.5T data that will be noisier than 3T (depending on the receive coil you use), but why do you say it didn't work very well? You need to visually inspect the surface accuracy to determine that.
cheers Bruce
On Tue, 13 Mar 2012, Charlotte Bernard wrote:
Thanks for your answer. But could you explain to me the aim of the white matter intensity normalisation. What is the incidence if this step doesn't work very well? In my subjects, we can see that the quality of this process is not very well! Maybe due to the acquisition on 1.5T?? Thanks,
Date: Mon, 12 Mar 2012 08:50:02 -0400 From: fischl@nmr.mgh.harvard.edu To: chakeba@live.fr CC: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] White matter intensity normalisation
Hi Charlotte
have you visually inspected them? I would be more concerned about the # of voxels in wm which is way too low, but it looks like you eroded the wm mask, so maybe it's ok. Take a look at the surfaces over the data in tkmedit and see if it appears accurate in subjects like #8. The WM intensity won't be exactly 110, and is usually below it due to partial volume effects.
cheers Bruce
On Mon, 12 Mar 2012, Charlotte Bernard wrote:
Dear Freesurfer users,
I run FS on my 670 subjects. All the scans were acquired on a 1.5T
Philips
scan. I noticed that my values in the WM are near 100-110 but not exactly 110.
I
think it is a problem! Is it link to the 1.5T? How can I fix it? I joined my log file of QA tools: could you help me to undertand the
output?
wm-anat-snr results:
SNR: Anatomical signal-to-noise ratio in white matter mWM: Mean white matter intensity value stdWM: Standard deviation of white matter intensity value voxWM: Total number of white matter voxels in eroded mask
SubjID SNR mWM stdWM voxWM 1 13,37* 104,18 7,79* 24181 <------ 2 14,79* 105,90 7,16* 22148 <------ 3 13,57* 104,75 7,72* 37341 <------ 4 12,67* 104,18 8,22* 29311 <------ 5 14,36* 105,44 7,34* 48083 <------ 6 11,40* 103,92 9,11* 47704 <------ 7 14,52* 104,71 7,21* 25706 <------ 8 9,92 101,42 10,22 40187 9 11,31* 104,28 9,22* 22967 <------ 10 11,61* 102,92 8,86* 19850 <------ 11 14,78* 105,05 7,11* 26607 <------ 12 11,36* 105,33 9,27* 15465 <------ 13 11,30* 103,69 9,18* 34685 <------ 14 11,73* 105,20 8,97* 40471 <------ 15 11,92* 104,04 8,73* 13777 <------ 16 9,46 102,04 10,79 30314 17 12,60* 106,40 8,44* 38432 <------
I thought it can help me to quickly identify the subjects who didn't
work
for this step. For example, the subject n°8 has a problem?
Thanks, Charlotte
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
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.
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl