The results of database comparisons will crucially depend on the parameters used for database creation.
▪When calculating the reference value for scaling the pixel values, diseased tissue should always be avoided. Therefore it might be helpful to have databases available with differing reference calculation, for example one using a gray matter mask, and another using a cerebellum mask.
▪Resolution mismatches between the database template and the processed subject image and inaccuracies in spatial normalization will result in edge artifacts in the z-score maps. This effect can be reduced by sufficient image smoothing. Again, it might be helpful to have databases available with different smoothing, to be applied depending on the size of the effect one is looking for.