We studied preparatory attention in patients suffering from frontotemporal dementia in the beginning stages of the disease, using an experimental test developed by LaBerge, Auclair, and Sieroff (2000). In this experimental test, a distracter can appear while subjects have to prepare to respond to a simple target. The probability that a distracter can appear in a trial is varied across three blocks. Normal controls show an increase of response times to targets (slope) as a function of the distracter probability:preparatory attention to the target is reduced by the increase of the distracter probability. Patients suffering from frontotemporal dementia show a slope of response times which is more than twice as large as the slope obtained by their matched controls. Such an abnormal increase of response times to targets is interpreted as a deficit in preparatory attention. Patients also show more omissions than controls. We suggest that this deficit in preparatory attention is related to the frontal lesion presented by the patients and can result in higher distractibility, a symptom frequently encountered in these patients.