Hi Daniel, there is no difference. Once you select the appropriate threshold, it does not matter whether it is a z-field, p-field, t-field, or -log10(p)-field because it gets binarized.
doug

On 9/8/13 9:33 PM, Yang, Daniel wrote:
Hi Doug,

Thanks! I didn't realize that the threshold is a negative exponent to a base of 10!

Just wondering, is there a webpage that I can read to understand why FreeSurfer chooses this "exponent" approach instead of z-value?

Also are 1.3, 2.3, 3.3, 4.3 two-sided, and 1, 2, 3, 4, one-sided? For example, when we specify configure overlay in Freeview?

Thanks!
Daniel

-- 
Yung-Jui "Daniel" Yang, PhD
Postdoctoral Researcher
Yale Child Study Center
New Haven, CT
(203) 737-5454

On 9/8/13 6:03 PM, "Douglas Greve" <greve@nmr.mgh.harvard.edu> wrote:

It is not a z-value. It is -log10(p), so -log10(.01) = 2
doug

On 9/7/13 9:13 PM, Yang, Daniel wrote:
Hi FreeSurfer Experts,

On this page (https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/QdecMultipleComparisons), it says: "In particular, thresholds of 1.3, 2, 2.3, 3, 3.3 and 4, corresponding to p-values of 0.05, 0.01, 0.005, 0.001, 0.0005 and 0.0001, which are common thresholds."

I am wondering how these thresholds were computed. 

In my own computation, I found z-value to be the following, but neither one-sided nor two-sided matches the thresholds of 1.3, …, 4 as shown above?

one-sided:
p-value=0.0500, z-value=1.6
p-value=0.0100, z-value=2.3
p-value=0.0050, z-value=2.6
p-value=0.0010, z-value=3.1
p-value=0.0005, z-value=3.3
p-value=0.0001, z-value=3.7

two-sided:
p-value=0.0500, z-value=2.0
p-value=0.0100, z-value=2.6
p-value=0.0050, z-value=2.8
p-value=0.0010, z-value=3.3
p-value=0.0005, z-value=3.5
p-value=0.0001, z-value=3.9

Thanks!
Daniel

-- 
Yung-Jui "Daniel" Yang, PhD
Postdoctoral Researcher
Yale Child Study Center
New Haven, CT
(203) 737-5454


_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer