Setayesh,
See
Kiebel, S and Poline, J and Friston, K and Holmes, A
and Worsley, K,
Robust smoothness estimation in statistical
parametric maps using standardized residuals from
the general linear model,
NeuroImage, 1999, 10:756-766
for details of how SPM99 estimates smoothness. In short, the inverse
of smoothness, roughness, is estimated as the sample variance of the
partial derivatives of the standardized residual images (that's the
data after subtracting off the signal and dividing by the residual
standard deviation).
I'm not sure I answered all of your question, though.
-Tom
On Wed, 18 Sep 2002, Barnden, Leighton (NWAHS) wrote:
> Dear all,
>
> Could someone please let me know how the smoothness of an image is
> determined? We have a template from a group of normal (SPECT) images. As
> such this template is smoother than each individual scan. To perform spatial
> normalization each scan needs to be smoothed first but the images, du to
> different uptake, have different noise level (/smoothness). One would then
> like to smooth each scan with a different kernel/fwhm (OK?) ... and hence
> rises the question of determination of the smoothness.
>
> Best wishes and thanks,
>
> Setayesh Behin-Ain
>
|