Pearls and pitfalls in human pharmacological models of migraine: 30 years' experience