On Tue, 29 May 2007 15:43:25 +0200, Marko Wilke <[log in to unmask]>
wrote:
Hi Michael,
I have never tried to estimate the smoothness of the deformations directly. However, you can
approximate the smoothness if you have analyzed modulated and unmodulated images by using
the estimated smoothness (which can be found at the bottom right on the results page):
FWHM_deformation = sqrt(FWHM_modulated^2 - FWHM_unmodulated^2)
The FWHM of the deformations is mostly influenced by the cut-off value (and also by the
regularization). If these parameters are not changed you only need to approximate the FWHM
once. The you can reorder the formula to:
FWHM_modulated = sqrt(FWM_deformation^2+FWHM_unmodulated^2)
You see that smoothness is added by the deformations and this usually leads to an increase of the
smoothness by 20-40%. However, this calculation was thought to make modulated and
unmodulated data comparable.
Best,
Christian
--
____________________________________________________________________________
Christian Gaser, Ph.D.
Assistant Professor of Computational Neuroscience
Department of Psychiatry
Friedrich-Schiller-University of Jena
Philosophenweg 3, D-07743 Jena, Germany
Tel: ++49-3641-935805 Fax: ++49-3641-935280
e-mail: [log in to unmask]
http://dbm.neuro.uni-jena.de
>Hi Michael,
>
>> I had a quick question about smoothing when doing VBM. I've read on
>> Christian Gaser's VBM website
>> (http://dbm.neuro.uni-jena.de/vbm/segmentation/) down at the bottom of
>> the page, that as a rule of thumb, smoothing should be 70% lower than
>> what your target smoothness is (given you left the default cut off for
>> spatial normalization at 25mm). So if for example, my target smoothness
>> was 12mm FWHM, then I should smooth my modulated gray matter images at
>> around 8-9mm.
>
>This only applies if you use MODULATED images as the modulation step
>will increase smoothness by virtue of the fact that the spatial
>normalization parameters (and hence, the Jacobian determinants of its
>matrix) are not calculated for every voxel. Therefore, for all those
>voxels "in between", interpolation is needed. If you check a j-image
>representing the Jacobian, you will be suprised of how few features you
>will be able to make out... therefore, the idea is that if you smooth
>unmodulated data by 12mm, you should smooth modulated data by about 9mm
>to get comparable results (I just ralized that I am repeating
>Christian's words here but there are worse sins :)
>
>> Is there anything out there published I can use to cite
>> in a paper for this "rule of thumb" or is this some esoteric thing that
>> people who use SPM have always done and just never bother to report it
>> in papers? I ask, because some of our reviewers want to know where is
>> the precedence for smoothing using this rule of thumb.
>
>Oh well, you could write that in order to account for the additional
>smoothness induced by the modulation step and in order to make the
>results comparable with earlier studies investigating gray matter
>density and due to the matched filter theorem, suggesting a width of
>about 12mm for cortical regions of interest, the final smoothing was
>chosen etc...
>
>... theoretically, one could assess the smoothness of the Jacobian and
>calculate its impact, and perhaps Christian knows if this has been done
>(which would be the reference you're looking for :)
>
>Best,
>Marko
>--
>===========================================================
==========
>Marko Wilke (Dr.med./M.D.)
> [log in to unmask]
>
>Universitäts-Kinderklinik University Children's Hospital
>Abt. III (Neuropädiatrie) Dept. III (Pediatric neurology)
> Hoppe-Seyler-Str. 1, D - 72076 Tübingen
>Tel.: (+49) 07071 29-83416 Fax: (+49) 07071 29-5473
>===========================================================
==========
|