Hi Kamen,
In FSL-VBM, brain size correction does not take into account the size of the skull. So indeed, it would compensate for a global shrinkage of the brains in elderly subjects (VBM is dedicated originally to see *local* changes in GM) and subsequently, you should not see any difference on the outer side of the cortex forming some kind of "rind"...
If you have used SPM2, the affine transformation between native GM and the template is taken into account by default in the modulation step (on the contrary to the default settings in FSL-VBM), so you are *not* globally compensating for a shrinkage of the brain (which is a very likely explanation for the differences between your two sets of results).
Cheers,
Gwenaelle
--- En date de : Mer 28.1.09, Kamen Tsvetanov <[log in to unmask]> a écrit :
> De: Kamen Tsvetanov <[log in to unmask]>
> Objet: Re: [FSL] WM detected as GM in fslvbm
> À: [log in to unmask]
> Date: Mercredi 28 Janvier 2009, 18h42
> Hi Steve,
>
> just two follow-up questions referring FNIRT and my
> results.
>
> We all know that the brain shrinks across the lifespan.
> Then we can hypothesize that the older brains are smaller
> and have thinner cortex than the young ones. One step in the
> FNIRT (fslvbm_2_template) algorithm corrects for brain size.
>
> Is it possible that this brain size correction does not
> take into account the size of the skull (like in SIENAX) and
> in fact "inflates" more than necessary the
> shrunken brains of the elderly, trying to match them to the
> size of the young healthy brains/template? Does it correct
> for cortical thickness as well?
> This could be a possible explanation why I get selection of
> significant voxel deeper in the cortex (at the border
> between GM and WM), instead areas at the outer site of the
> cortex (seen in my SPM results).
>
> I appreciate your opinion.
>
> Best.
>
>
> ----- Original Message ----
> From: Steve Smith <[log in to unmask]>
> To: [log in to unmask]
> Sent: Sunday, January 25, 2009 1:16:30 PM
> Subject: Re: [FSL] WM detected as GM in fslvbm
>
> It is possible to get a difference in GM within regions
> that are on average (across subjects) more WM than GM,
> because they can still contain some "GM",
> particularly when you have applied smoothing to the data. I
> would overlay this on your mean GM map that the scripts
> generated to reassure yourself, but they may be ok.
>
> Cheers.
>
>
> On 24 Jan 2009, at 14:10, Kosta Markov wrote:
>
> > Dear FSL users,
> >
> >
> > I am trying to use FSL-VBM v1.1 on two groups of
> subjects (young vs old) to
> > look at effects of ageing in GM. The VBM output
> revealed some strange results -
> > highly significant clusters of voxels (showing
> decrease of GM with age)
> > within white matter areas (attached snapshot). Do you
> have any idea what
> > could possibly be the reason for these ambiguous
> results?
> >
> > Here are the steps I have done:
> >
> > - individual BET with varying -f and -c option to
> provide proper brain
> > extraction
> > - fslvbm_2_template -n
> > - fslvbm_3_proc 0.2
> > - randomise -i GM_mod_merg_s4 -o GM_mod_merg_s4
> -GM_mask -d design.mat -t
> > design.con -n 5000 -tfce(-c 3) -V
> > - the design was created in the Glm wizard with two
> contrasts (A>B: 1 -1;
> > B>A:-1 1)
> > - the attached snapshot shows the
> GM_mod_merg_s4_tfce_corrp_tstats1
> > (intensity threshold 0.999-1) overlaid on the standart
> MNI152_2mm
> >
> > Any suggestions are greatly appreciated.
> >
> > thanks,
> > Kosta
> >
> > <p_value-0.001.jpg>
>
>
> ---------------------------------------------------------------------------
> Stephen M. Smith, Professor of Biomedical Engineering
> Associate Director, Oxford University FMRIB Centre
>
> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
> +44 (0) 1865 222726 (fax 222717)
> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
> ---------------------------------------------------------------------------
|
