Temporal context effects in pigeons' memory for event duration

Abstract These experiments extend previous studies of how pigeons' memory for the duration of events is affected by the length of the delay and intertrial intervals. Pigeons were trained using a delayed matching-to-sample procedure to peck one color following a long sample and another color following a short sample. Experiments 1a and 1b tested a prediction, derived from the relative duration hypothesis, that systematic errors in response to delay manipulations should be greater in the context of a short intertrial interval than in the context of a long one. Pigeons were given delay tests in the context of either 10-s or 60-s intertrial intervals. Systematic errors in response to delay manipulations were of greater magnitude in the context of the 10-s intertrial interval, as predicted by the relative duration hypothesis. Experiments 2a and 2b examined whether stable delay and intertrial intervals during training are important for the development of a temporal reference memory. Variability in the intertrial or delay interval during training had only minor effects on acquisition and did not affect pigeons' tendency to make systematic errors in response to subsequent manipulations of these intervals. Moreover, even during acquisition, birds trained with variable delays showed a systematic increase in the tendency to make choices appropriate to the short sample as a function of the training delay, with the least bias occurring at an intermediate delay. Thus, reference memory appeared to be based on an average of the delays experienced during training.