Hi Koene,
I tried it with spm8 and it worked if you also specify the last
argument, e.g.:
CorrClusTh(SPM,0.005,0.05,[1:500])
Possibly you need to change the value 500 to a larger one in case 500
voxels in a cluster are not enough to control for false positives at the
desired level.
Cheers,
Thilo
On Wed, 2011-07-20 at 20:10 +0100, Koene Van Dijk wrote:
> Hi there,
>
> I'm trying to use CorrClusTh.m that I downloaded here:
> http://www2.warwick.ac.uk/fac/sci/statistics/staff/academic-research/nichols/scripts/spm/johnsgems5/CorrClusTh.m
>
> which has the following details regarding version:
> $Id: CorrClusTh.m,v 1.12 2008/06/10 19:03:13 nichols Exp $ Thomas Nichols, Marko Wilke
>
>
> On Tom Nichols' page it says that this one will work with SPM5. I am currently using SPM8 and was wondering if I can use the same script. It seems not because I get the following error:
>
> ===========================================
> >> CorrClusTh(SPM,0.005,0.05)
>
> Search Volume: 902629 2.00x2.00x2.00 voxels
> 7221032.00 cmm, 7.22 L, 16608.93 RESELS
> 7.99x7.96x6.63 mm FWHM, 3.99x3.98x3.31 vox FWHM
>
> ??? Error using ==> spm_P
> Too many output arguments.
>
> Error in ==> CorrClusTh at 84
> [P Pn Em En EN] = spm_P(1,0,u,df,STAT,R,n,S);
> ===========================================
>
>
>
> Has anybody used CorrClusTh.m with SPM8? Or does anybody know how I can adapt the script to work with SPM8?
>
> Any help is appreciated a lot!
>
> Best,
> Koene
|