We studied, in healthy elderly subjects (aged from 63 to 83 years) and adults (aged from 20 to 32 years), ocular saccades in two conditions: one the one hand, the gap condition, where the central target disappears; then follows a period of 200 ms during which the fixation and attention were disengaged; finally, a visual target appears in the periphery. On the other hand, the overlap condition, in which the peripheral target appears when the central target is still present, the subject should voluntarily disengage his attention and fixation to orient them toward the peripheral target. These paradigms stimulate automatic versus controlled triggering of saccades. The average saccade latency (measured by video-oculography) was longer in the elderly, and irrespectively of the condition. However, the elderly as the young subjects produced shorter latencies in the gap condition than in the overlap condition. Moreover, in the gap condition, we observed the emergence of a considerable number of reflex saccades with very short latency (between 80 and 120 ms, minimal conduction time) called "express saccades". The occurrence rate of such saccades was similar in the young and the elderly subjects. These results suggest the existence of separate circuits, one non-being sensitive to age (express saccades), the other suffering the effects of aging (controlled saccades). In another ongoing study, this methodology has been applied to patients with Lewy body dementia. The preliminary results from three patients showed an abnormal slowness of latencies, even in the gap condition expected to promote automatic and reflex saccades. Furthermore, we observed a total absence of saccades with express latency. These promising results suggest a deficit even for automatic and express saccades in these patients.