About 3dClusterSim Correction

Submitted by wleewell on Wed, 08/08/2012 - 19:29

严老师好！
想请问下新版的REST会加入AFNI中的3dClusterSim校正吗，最近听一位老师说3dClusterSim要比Alphasim correction更好，目前更推荐，严老师可否讲一下这两个方法的特点及对比呢？
谢谢！

Submitted by Qing Zhao on Wed, 08/08/2012 - 22:32

The following text is just a copy of a part of the help information of AFNI's 3dClustsim

"This program is like running AlphaSim once for each '-pthr' value and then
  extracting the relevant information from its 'Alpha' output column.
 ++ One reason for 3dClustSim to be used in place of AlphaSim is that it will
    be faster than running AlphaSim multiple times.
 ++ Another reason is that the resulting table can be stored in an AFNI
    dataset's header, and used in the AFNI Clusterize GUI to see estimated
    cluster significance (alpha) levels."

Thus, I think their algorithm are same, 3dClustsim is just a streamlined form.
you can see <http://afni.nimh.nih.gov/pub/dist/doc/program_help/3dClustSim.html> for detail

Submitted by YAN Chao-Gan on Thu, 08/09/2012 - 00:00

同意。算法是一样的。只不过前者会直接出来很多个不同校正前阈值的所需团块大小。
新版会加入一个类似FSL的easythresh的校正。