该FDR校正的方法是否存在问题？

Submitted by 朝阳七号 on Mon, 07/30/2012 - 15:15

老师您好：
       我现在虽然在做任务态的实验，想请教您关于FDR校正的问题。我的实验是比较一组被试学习前、后脑区激活的变化，关心的是枕叶的激活。学习前单样本T检验得到激活簇A，学习后单样本T检验得到激活簇B，配对T检验后得到激活簇C。可是三个检验结果都经不起全脑或枕叶的FDR校正（即用全脑做mask或枕叶做mask，Q值取0.05）。我可不可以采用以下方法：先做P=0.05未校正，再在rest中save clusters ，分别保存P=0.05未校正的结果（A、B、C）作为各FDR检验的mask？这样的话结果经得起校正。

我暂时没找到关于此校正方法的参考文献，在此求证该方法是否合理，希望老师能够给予指导！感激不尽O(∩_∩)O~

Forums: 
Resting-state fMRI topics

AlphaSim in Rest fMRI

Submitted by modi_shilpi on Mon, 07/30/2012 - 15:12
I performed an fMRI task for attention. Then using the 'con files' of all the subjects, I performed a multiple regression analysis in SPM8, with a behavioral score as covariate to identify the regions that have a positive or negative correlation with the score. I want to perform AlphaSim for this multiple regression analysis to find the minimum cluster size for alpha 0.05. My uncorrected p value is 0.01, slice thickness is 3 mm.
 
What should I input for the mask image? Can I use the 'mask' image generated during the regression analysis?
Forums: 
Other