Hey Sandra
Disclaimer: not an SPM expert, just a user.
I tested this very thing myself and came to the conclusion that the REST alpha sim probably hasn't undergone the same 'fix' as AFNIs tool. I ran them both a bunch of times and found they were typically about 10-15 voxels different in what they considered significant at p<.05. Also I found that this difference basically disappeared when I ran AFNI 3dclustsim using the -nopad flag (which turns off the 'fix')
Why not just use 3dclustsim? It just requires a mask as an input in order to define a space to do the simulations. So why can't you just use 3Dclustsim to correct results otherwise produced in SPM? (Perhaps someone else can chime in here if they know better)
Sent from my iPhone
> On 25 Jul 2016, at 6:53 PM, Sandra <[log in to unmask]> wrote:
>
> Dear SPM experts,
>
> does anyone know whether the bug-fix as it's done for the 3dClustSim by the AFNI developers (for version after May 2015) was also implemented in the SPM-compatible alphasim REST toolbox implementation or whether the latter rather mirrors the older 3dClustSim version? In that case, is there any reliable alternative SPM toolbox for cluster-level FWE control?
>
> Thank you very much in advance,
> Sandra
|