The primary headaches, migraine with (MA) and without aura (MO) and cluster headache, all carry a substantial genetic liability. Familial hemiplegic migraine (FHM), an autosomal dominant mendelian disorder classified as a subtype of MA, is due to mutations in genes encoding neural channel subunits. MA/MO are considered multifactorial genetic disorders, and FHM has been proposed as a model for migraine aetiology. However, a review of the genetic studies suggests that the FHM genes are not involved in the typical migraines and that FHM should be considered as a syndromic migraine rather than a subtype of MA. Adopting the concept of syndromic migraine could be useful in understanding migraine pathogenesis. We hypothesise that epigenetic mechanisms play an important role in headache pathogenesis. A behavioural model is proposed, whereby the primary headaches are construed as behaviours, not symptoms, evolutionarily conserved for their adaptive value and engendered out of a genetic repertoire by a network of pattern generators present in the brain and signalling homeostatic imbalance. This behavioural model could be incorporated into migraine genetic research.