The ephemeral reward task involves presenting subjects with a choice between two stimuli that each provide an identical reward. If one stimulus (stimulus A) is selected, reinforcement associated with that stimulus is provided and the trial ends; however, if the other stimulus (stimulus B) is selected, the reinforcer associated with that stimulus is provided and the subject is also allowed to obtain the reinforcer associated with stimulus A. Thus, it is optimal for the subject to choose stimulus B. Although wrasse (a fish) and parrots are able to quickly learn to maximize reinforcement, other species (including humans, nonhuman primates, rats, and pigeons) struggle to choose optimally. Research with pigeons and rats found that insertion of a delay between choice and initial reinforcement facilitates optimal performance, suggesting that mitigating impulsivity may increase optimal choice; however, research with human participants does not point to a relationship between measured and self-reported impulsivity and suboptimal performance on the task. For this reason, it was posited that subjects sometimes failed to discriminate the relation between choice and the second reinforcer. To test this hypothesis, we proposed that if the time between trials was increased, the relation between the subject’s choice and the second reinforcer would be more discriminable and optimal choice might be achieved. We found, however, that increasing the intertrial interval from 5 s to 30 s did not facilitate optimal choice. We conclude that the earlier effect of increasing the delay between choice and the initial outcome probably reduced impulsive choice, at least for pigeons and rats.

短暂奖励任务涉及让受试者在两种刺激之间进行选择，每种刺激都提供相同的奖励。如果选择了一种刺激（刺激 A），则提供与该刺激相关的强化，并且试验结束；然而，如果选择其他刺激（刺激 B），则提供与该刺激相关的强化物，并且还允许受试者获得与刺激 A 相关的强化物。因此，受试者选择刺激 B 是最佳选择。虽然濑鱼（鱼）和鹦鹉能够快速学会最大化强化，但其他物种（包括人类、非人类灵长类动物、大鼠和鸽子）很难做出最佳选择。对鸽子和老鼠的研究发现，在选择和初始强化之间插入延迟有助于实现最佳表现，表明减轻冲动可能会增加最佳选择；然而，对人类参与者的研究并没有指出测量和自我报告的冲动与任务表现不佳之间存在关系。因此，假设受试者有时无法区分选择和第二个强化物之间的关系。为了检验这一假设，我们提出，如果增加试验之间的时间，受试者的选择和第二个强化物之间的关系将更具可辨别性，并且可能会实现最佳选择。然而，我们发现，将试验间隔从 5 秒增加到 30 秒并不能促进最优选择。我们的结论是，增加选择和初始结果之间的延迟的早期效应可能会减少冲动选择，至少对于鸽子和老鼠来说是这样。