Dear Experts,
I have run two VBM models using randomise and tfce, one model with a variable of interest plus a nuisance regressor, and one model without the nuisance regressor. The results from the model with the nuisance regressor omits some of the voxels that were in the results from the model where this wasn't included. I conclude from this that the nuisance regressor mediates the relationship between my variable of interest and grey matter in those voxels that are no longer statistically significant. Is this conclusion valid, or do I need to run some sort of test to determine if the two clusters differ spatially? If the latter is the case, how do I go about doing this please?
Thanks.
GD
|