Pearls and pitfalls in experimental models of spreading depression