The Frontal Assessment Battery (FAB) is a short neuropsychological tool aiming to assess executive functions at the bedside. Two-hundred and thirty-six normal controls were administered the FAB and three other tasks assessing attentional and executive functions. The FAB was also administered to 28 patients suffering from mild Alzheimer's disease (AD, n. 15 subjects) or frontotemporal dementia (FTD, n. 13 subjects). The FAB showed good concurrent and discriminant validity and high internal consistency. Test-retest and inter-rater reliability were fairly good. A multiple regression analysis showed a significant positive effect of education and a negative effect of age. Cut-off values of non-parametric distribution were computed. A difference on FAB scores within dementia patients was observed, with subjects with FTD performing worse than patients suffering from AD.