No, I've found that the neurals with those, and some of the other categories, can sometimes be too puzzling. I suppose I could apply some filters to the rules but in the meantime I'll concentrate on the current method as shown on this thread.