The role of 'nothing' in memory for event duration in pigeons

Abstract

The research reported herein is designed to test a signal detection account of the choose-short effect, which is the tendency of birds to report (after a long delay) that a short- duration sample was presented, regardless of whether a short or long sample initiated the trial. According to the detection account, the choose-short effect arises because birds learn to selectively search memory for evidence that the long sample appeared on a given trial. This idea is tested, in part, by replacing short-sample trials with no-sample trials and showing that performance is unaffected by this manipulation for birds exhibiting a choose-short effect. In addition, when no samples and long samples were associated with the same choice alternative (and the short sample was associated with the other alternative), birds were flexible enough to learn to respond on the basis of the presence versus the absence of the short sample (and, as a result, a choose-long effect was observed).