Hi,
I try to apply cluster to p-values from randomise, passing along a tstat as cope.
Having one big cluster with multiple blobs I am more interested in local maxima.
As local maxima of p-value is not really of interest I used fslmath to mask the tstat file with the p-value thresholded.
The image is exactly what I want.
Then I ran cluster with this map thresholded and with olmax.
cluster -v -i cope3_main_fx_tstat1_corrpthresh.05.nii.gz -t .95 --mm --olmax=olmax_test.txt
There are really a few maxima as compared to what I can see in the image, either with 6 or 26 connectivity (same results).
Any idea? Your help is very welcome.
Thanks.
Basile
|