Unfortunately, the MRIQuestions.com forum became unmanageable due to spam attacks. I recommend an alternative resource: www.imagingQA.com

Hello All,
I am analyzing the brain activation regions by means of SPM12. Recently I have applied the algorithm FWE. Thanks to that the results are much more accurate. Nevertheless I have doubts how I should interpret the table with statistics which appears bellow glass brain.

At first, I setup significance level = 0,001
Next, I processed image.
Then, I estimated the number of false activated voxels in cluster.
In order to do that I calculated
p(FWE+corr) * ke
As far as I know:
p(FWE+corr) - the probability of false activation in cluster
ke - the number of voxels in cluster

I calculated p(FWE+corr) * ke for each cluster. It allows me to calculate the average probability of false activation for all identified clusters.
This vale equals 10-7.


I am surprised that I have only asked for significance level = 0,001, the result is much more accurate (significance level = 0,0000001).

I would like to ask You for comments for the following ideas:
a) Am I correct with interpretation of p(FWE+corr) - is it the probability of false activation in cluster?
b) the FWE algorithm seems to behave greedy, it achieves the best fit regardless of significance level = 0,001 assigned in setup.

Regards, Anna