Hello Shruti,
The values in the tfce_corrp images contain corrected p ( actually 1-p ) values. The cluster command is used to only report the subset of voxels passing a certain significance threshold, in this case an alpha of 0.05 ( which corresponds to a 1-p of 0.95 ).
Kind Regards
Matthew
> On 24 Jan 2017, at 11:13, Shruti Narasimham <[log in to unmask]> wrote:
>
> Dear all,
>
> I have a query regarding the 0.95 threshold in the cluster command.
>
> I read and was told previously that the dr_stage3_tfce_corrp results show significant functional connectivity differences for my groups A>B, C<D, etc. I don't understand that why then in the cluster command, do we supply a further threshold of 0.95 to get significant results from the stage3_tfce_corrp blobs. Does this mean that not all the tfce_corrp results are significant?
>
> Could someone please explain this?
>
> Thank you.
>
>
> Regards,
> Shruti
|