Objective: Patients with visual extinction have difficulty detecting a single contralesional stimulus when a second stimulus is simultaneously presented on the ipsilesional side. The rarely reported phenomenon of visual anti-extinction describes the opposite behavior, in which patients show greater difficulty in reporting a stimulus presented in isolation than they do in reporting 2 simultaneously presented stimuli. S. J. Goodrich and R. Ward (1997, Anti-extinction following unilateral parietal damage, Cognitive Neuropsychology, Vol. 14, pp. 595–612) suggested that visual anti-extinction is the result of a task-specific mechanism in which processing of the ipsilesional stimulus facilitates responses to the contralesional stimulus; in contrast, G. W. Humphreys, M. J. Riddoch, G. Nys, and D. Heinke (2002, Transient binding by time: Neuropsychological evidence from anti-extinction, Cognitive Neuropsychology, Vol. 19, pp. 361–380) suggested that temporal binding groups contralesional and ipsilesional stimuli together at brief exposure durations. Method: We investigated extinction and anti-extinction phenomena in 3 brain-damaged patients using an extinction paradigm in which the stimulus exposure duration was systematically manipulated. Results: Two patients showed both extinction and anti-extinction depending on the exposure duration of stimuli. Data confirmed the crucial role of duration in modulating the effect of extinction and anti-extinction. However, contrary to Humphreys and colleagues' (2002) single case, our patients showed extinction for short and anti-extinction for long exposure durations, suggesting that different mechanisms might underlie our patients' pattern of data. Conclusion: We discuss a novel “attentional waiting” hypothesis, which proposes that anti-extinction may be observed in patients showing extinction if the exposure duration of stimuli is increased.