Hello Dan,
It is not really appropriate to use the tool cluster in this case - the GRF assumptions in cluster do not apply to TFCE values. If you wish to obtain clusters with discrete p-values from randomise, then the -c option is the one to use.
Hope this helps,
Kind regards
Matthew
> Hello,
>
> Apologies if this has been addressed before, but I couldn't find any related posts.
>
> I'm trying to extract the cluster probability value for individual clusters from a randomise analysis (tfce_corrp_tstat image). I'm able to get some of the statistics using the fsl "cluster" command, but not the cluster statistics. I've tried playing around with the GRF flags, but honestly I'm not sure what values to select for smoothness (d) or volume, and each attempt gives me NaN values for p and log(p).
>
> Any help will be greatly appreciated!
>
> Thanks,
>
> Dan
|