The aim of the study was to estimate the sensitivity of a brief self-paced visual search task to increased levels of sleepiness as a consequence of 40 h of sleep deprivation. Time-of-day effects on this task, on subjective sleepiness and on oculomotor performance changes, were also assessed. Eight normal subjects slept for three nights in the laboratory (adaptation, baseline, recovery). Baseline and recovery nights were separated by a period of 40 h of continuous wakefulness, during which subjects were tested every 2 h from 10:00 to 22:00 h on both days preceding and following the sleep deprivation night, as well as from 24:00 to 08:00 h during the deprivation period. At the same time, subjects filled in a visual analogue sleepiness scale and the Stanford Sleepiness Scale (SSS). As regards cognitive performance, significant effects were found on speed measures, while accuracy was not affected. The number of explored rows was higher after the baseline night than after the sleepless night, and showed a consistent time-of-day trend. Omissions ratio (OR), false positives ratio (FPR) and hits ratio (HR) did not show any significant effect. Subjective ratings of sleepiness varied according to speed measures, being affected by sleep deprivation and time of day. Since similar effects were found with an oculomotor task, detrended functions for all variables across the 40 h of continuous wakefulness were calculated. A circadian effect was found, in which speed measures seem to be more affected than accuracy ones in both visual search and oculomotor tasks. It is concluded that 40 h of prolonged wakefulness significantly impairs performance in a brief cognitive visual search task. Such a performance worsening is evident on speed, but not on accuracy indices, and is strictly related to the deterioration of oculomotor performance, indicating a clear circadian effect.