Dear Christa,
I think you have misunderstood.
The values in the corrp images are p-values (corrected for multiple comparisons) and do not correspond to this critical value. The critical value applies to the raw statistic values used to obtain the p-values. In general I would not worry about the critical value and just look at the corrp values above 0.95.
All the best,
Mark
On 22 Apr 2013, at 21:57, Christa Watson <[log in to unmask]> wrote:
> Hello,
> I have just discovered the critical threshold number as a randomise output and have read the posts about what it means, however I still have a few of questions. I will use my data as an example.
> I ran an analysis and got the message:
> "Critical Value for: *_tfce_corrp_tstat1 is: 202965"
>
> Previously, you have said that this is the value to use if I would like to threshold the image and that you would expect this value to be in the range of a *_tfce_corrp_tstat image. I am correct in summarizing your previous statements?
>
> What I am confused about is that I see only pvalues when I look at the *_tfce_corrp_tstat1.
> How could I view the image to see a range that would contain this number?
> Is a pvalue of .95 (reversed as in fslview) equivalent to the tfce threshold of 202965 for my data?
>
> Second, if I want to threshold the image instead of using fslmaths *_tfce_corrp_tstat1 -thr 0.95 should I use fslmaths *_tfce_corrp_tstat1 -thr 202965 ?
>
> Thank you for your help,
> Christa
|