Dear Adrian,
for VBM type analyses, standard SPM cluster size correction should not be
reported because of the spatial non-stationarity of error. You may want to
try the results reports from Christian Gasers VBM5 toolbox, but they are
not yet properly integrated in SPM5 - Christian and me are trying to find
a solution for this problem.
Volkmar
On Thu, 4 Oct 2007, Adrian Imfeld wrote:
> Hi SPM users,
>
> I'm analyzing DTI data with SPM5 and I'd like to do group comparisons
> (n1=13, n2=13) of fractional anisotropy images and within group correlations
> of anisotropy with a performance score. I'm trying different smoothing
> kernels (4x4x4mm vs. 8x8x8mm) and different alpha-levels (p<0.01 vs. p<0.05,
> uncorrected).
>
> Question 1: After inspection of the SPM5 statistical tables, I'm inclined to
> use smoothing with 8x8x8mm and an alpha level of 0.05. Do you think this is
> reasonable or would you use other parameters?
>
> Question 2: The size of clusters seems more important to me than their
> Z-value peek voxel. Is it okay to focus on the cluster-level analysis
> (p_corrected column) instead of the voxel-level analysis?
>
> Any help is highly appreciated.
>
> Regards,
>
> Adrian Imfeld
>
--
Volkmar Glauche
-
Department of Neurology [log in to unmask]
Universitaetsklinikum Freiburg Phone 49(0)761-270-5331
Breisacher Str. 64 Fax 49(0)761-270-5416
79106 Freiburg http://fbi.uniklinik-freiburg.de/
|