FDR is based on the data, so you can't really say that a voxel-wise threshold is low or high. But 1x10^04 is actually not too bad, considering that the bonferonni thresh would be about 3x10^-7. The jump you see is not entirely unexpected, but I'm going to have to refer you to the FDR liturature (or Tom Nichols:) to get a deeper explanation.
The new tools that I am working on will use clustering to help reduce the multiple correction penalty. Should be ready in the next couple of weeks.
doug
Sasha Wolosin wrote:
When I set FDR = .05, or anything under .85, and select "set threshold using FDR" the min threshold is always set to 4.03, which is about p= .0001, correct? Would you normally require such low p-values for this type of analysis? Why would such varied FDRs result in the same min threshold? We do not see any results when the min>2.0 (p=.01), so this is very problematic for us. Also, at just over FDR = .85 (I realize this is absurdly high) the min threshold jumps down to 1.23, do you have any idea why this jump might occur? Thanks, Sasha
Bruce Fischl fischl@nmr.mgh.harvard.edu >>>
.05 is fairly standard I think. On Mon, 9 Jan 2006, Sasha Wolosin wrote:
Is this something you apply frequently at MGH? What FDR value would
you
typically require?
Doug Greve greve@nmr.mgh.harvard.edu >>>
Currently, mris_glm does not attempt to correct for multiple comparisons, though this feature will be in the next release. In the mean time, you can correct for multiple comparisons using FDR in tksurfer.
Sasha Wolosin wrote:
Dear all, I would like to learn more about group analysis using mris_glm. How does the program correct for the problem of multiple comparisons? Is there any way to determine the rate of false positives? Thanks, Sasha
Sasha Wolosin Research Assistant Developmental Cognitive Neurology Kennedy Krieger Institute 707 N. Broadway Baltimore, MD 21205 ph: (443) 923-9270 wolosin@kennedykrieger.org
Disclaimer: The materials in this e-mail are private and may contain Protected
Health Information. Please note that e-mail is not necessarily confidential or secure. Your use of e-mail constitutes your acknowledgment of these confidentiality and security limitations. If
you
are not the intended recipient, be advised that any unauthorized use, disclosure, copying, distribution, or the taking of any action in reliance on the contents of this information is strictly prohibited.
If
you have received this e-mail in error, please immediately notify the sender via telephone or return e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer