Rats were trained on a novel attentional task adapted from the five-choice serial reaction time test first developed by Carli et al. [1983; Behav. Brain. Res., 9, 361–80]. The novel task required rats to detect the occurrence of brief light flashes in one of two spatial locations following trial initiation by a lever press. Blank trials were interleaved with stimulus trials and the rat had to make a different response to indicate the absence of a light. Occasional light-alone trials were also presented in which the visual stimulus appeared without forewarning. Pre-exposure to food for 60 min prior to test increased response omissions for all trials and slowed correct response latency, but failed to significantly alter accuracy. A decrease in light-stimulus duration (1–0.4 s) decreased accuracy, increased the reporting of blank trials and the number of light-alone trial omissions. Scopolamine (0.03–0.1 mg/kg) and scopolamine methylbromide (0.1 mg/kg) failed to affect accuracy, but increased light-alone trial omissions and lengthened correct response latency. The results confirm that the novel task is able to distinguish between motivational and attentional manipulations, and imply that scopolamine affected performance via nonattentional mechanisms.