Hello,
I am trying to run 2-sample t-test equivalent to test for differences in voxel-wise correlation coefficients using randomise. For one group I am able to run randomise without any problem but as soon as I try with two groups I receive the message:
Warning: tfce has detected a large number of integral steps.
I looked at the post from 2010 and have double checked the model and the data. The number of subjects matches the model and the data are not corrupt. I don't know what else to try, and like in the previous post, randomise will run for hours/days but I don't get any output. I haven't tried to see if it will actually stop because after 2 days I assumed it had frozen. Is there some other reason that I could be overlooking for why this problem is occurring other than: 1) matching the order and number of the images and the design matrix, 2) having corrupt data?
Thanks,
Virginia
|