The classification and nomenclature of motor neuron disease, whether sporad
ic or familial, is confused. For example, both the sporadic and familial mo
tor neuron diseases are phenotypically heterogeneous and, in familial ALs,
phenotypic heterogeneity correlates only weakly with different underlying m
utations in the SOD1 gene. We propose a classification which is based on un
derlying causative mechanisms, where these are known, but which also recogn
izes different clinical phenotypes when the cause is unknown, This classifi
cation is flexible, and allows re-attribution of clinical syndromes when th
eir causation is understood. Currently uncertain associations - for example
, a possible association of ALS with cancer - are given tentative recogniti
on in this classification. In addition, this new classification recognizes
geographical clustering and descriptions of unusual motor neuron disorder p
henotypes of unknown origin in different parts of the world.