The quality of the neuropsychological investigation of behavioural changes in frontal dementias is dependent on the specificity of the behavioural analyses. A review of methodological issues and published findings identifies possible potential reasons for misinterpretation of research neuropsychological data in these studies. The areas reviewed include the operational definitions of terms such as ''frontal lobes'' and ''frontal functions''; identification of experimental limitations in frontal lobe research; and highlights of neuropsychological investigations of frontal lobe functions. These deliberations suggest a practical approach to the behavioural assessment of frontal dementias.