Shimp C P
J Exp Anal Behav. 1981 Nov;36(3):303-15. doi: 10.1901/jeab.1981.36-303.
A procedure was developed to enable nonverbal organisms to report what they remember of the temporal organization of their recent behavior. A baseline behavior with known temporal structure was established by a concurrent variable-interval variable-interval schedule for two temporal patterns of behavior (two different classes of reinforced inter-response times). The five pigeon subjects emitted these two temporal patterns on a center key and were occasionally given a short-term memory probe for their most-recently-emitted pattern. The probes consisted of symbolic delayed matching-to-sample tests, in which a response on a green side key was reinforced if the most recent pattern belonged to the shorter reinforced class, and a response to a red side key was reinforced if the most recent pattern belonged to the longer reinforced class. All subjects could report with over ninety percent accuracy what their most recently emitted behavioral pattern was when a retention interval separating the pattern from the memory probe was only .1 seconds. The retention interval was then manipulated, and it was found that recall for a pattern was frequently above chance after a delay of as much as eight seconds. Thus, pigeons can remember their most recent interresponse time not only right after it is emitted, but for several seconds thereafter. In other conditions, the patterns themselves were manipulated. It was found that as the patterns became more similar, discrimination became poorer. These results agree with the view that reinforcement tends to organize and integrate the local structure of behavior to the extent to which that structure is remembered.
开发了一种程序,以使非语言生物能够报告它们对近期行为时间组织的记忆。通过对两种行为时间模式(两种不同类别的强化反应间隔时间)采用并发可变间隔可变间隔时间表,建立了具有已知时间结构的基线行为。五只鸽子受试者在中心键上发出这两种时间模式,并偶尔针对它们最近发出的模式进行短期记忆探测。探测由符号延迟样本匹配测试组成,如果最近的模式属于较短的强化类别,则对绿色侧键的反应得到强化;如果最近的模式属于较长的强化类别,则对红色侧键的反应得到强化。当将模式与记忆探测分开的保持间隔仅为0.1秒时,所有受试者都能以超过90%的准确率报告它们最近发出的行为模式是什么。然后对保持间隔进行操纵,发现即使延迟长达八秒,对模式的回忆也经常高于随机水平。因此,鸽子不仅能在发出最近的反应间隔时间后立即记住它,而且在之后的几秒钟内也能记住。在其他条件下,对模式本身进行了操纵。发现随着模式变得更加相似,辨别能力变得更差。这些结果与以下观点一致,即强化倾向于在行为的局部结构被记住的程度上对其进行组织和整合。