Hello,
I am comparing SPECT data of 100 Alzheimer patients. They have been divided into groups of converted, questionable, and normal patients, with a MannWhitney test run on each group. Now I must find the maximum threshold for comparison for a p value of 0.005. I have tried running AlphaSim simulations a cluster connectivity radius of 1, gaussian filter width 10, 1000 iterations, and a mask of one of the patients.
This however has not given me the desired results. A reliable source has suggested the threshold value will be 2.798 with a maximum cluster size of 692. I will need to remove all data from the functional datasets that is below the necessary cluster size, so I will be running 3dmerge on each of the three MannWhitney tests, but I am still not sure how to acquire the correct threshold value for my tests.
Please let me know if you can help.
Thanks,
Ellen