Hi,
Yes, I set the rate to various values to test it, and all resulted in the same threshold. I didn't mask to brain, because that causes an error message (it cites the lack of a brainmap.mgz as the reason, so that at least I know is due to it being processed in an older stream). So I wonder if FDR requires that new processing as well, in stable3. By the way, these are the commands I used for view the data:
setenv SUBJECTS_DIR /space/ventzl/1/users/SUBJECTS_DIR source /usr/local/freesurfer/nmr-stable3-env cd $SUBJECTS_DIR tkmedit-sess -analysis EMerror -contrast ASevASc -space tal -map sig -s BwGroupError5 -isxavg random -fthresh 1.3
Thanks, Rob
Doug Greve wrote:
what do you mean by "staring p value"? Do you mean the FDR itself?
Robert Levy wrote:
Hi,
Is it possible to use FDR in stable3 for data that was processed prior to the introduction of brainmask.mgz? It seems like I can view the data just fine in stable3, or stable2 for that matter (though not stable 4), but FDR gives results that seem wrong. FDR sets the threshold to -log10(p)=4.279537 [p=.00005] no matter what the starting pvalue is. Thanks, Rob