sdm-help-list > Very high t values transformed from original z values.
Showing 1-2 of 2 posts
Display:
Results per page:
Jun 22, 2024  11:06 AM | Runchen Gan
Very high t values transformed from original z values.

Hello,


I'm facing a strange problem. In some early study using PET of fMRI, the authors only reported z values and have few subjects. When I usd the website of SDM to transform z to t, I found the some t values were extremely high (even over 200 or reached inf). I found a similar question in the past forums, but still don't get a good solution way.


Now, I tried 2 ways to solve these problems, but I didn't know whether it works.


1) I have double checked the values authors reported. For inf values, I use a value a little higher value than the max value (e.g. the max value is 7.8, I would use 8 to replace the inf)


2) For very high t values, I would like to use the p values, and use p transform to t values. But I'm not sure the feasibility of this approach.


Thanks for reading this. I hope this problem could be resolved. 


Runchen

Oct 13, 2024  03:10 AM | xianyu wang
RE: Very high t values transformed from original z values.

Hello! I also encountered a certain coordinate with a voxel z value greater than 9.99 when extracting coordinates. After converting it to t-value, it displays as inf. How can I solve this situation? (It does not report the p-value of voxels, only the p-value of clusters)