RETROSPECTIVE AND PROSPECTIVE SHORT-TERM-MEMORY IN DELAYED-RESPONSE TASKS IN RATS

Separate groups of rats were trained and tested on asymmetrically and symmetrically reinforced successive delayed matching-to-sample (DMTS) or delayed discrimination (DD) tasks in Experiment 1. Each rat received training and testing on symmetrically reinforced DMTS and DD tasks in Experiment 2. The only difference between each task was that the rats had to respond correctly to a light or tone test stimulus, S2, if it matched a light or tone sample stimulus, S1, in DMTS, but could respond to either S2 if S1 had been a particular stimulus in DD. Only correct lever presses were reinforced in the asymmetrically reinforced version of each task. Both correct presses and correct omissions were reinforced in the symmetrically reinforced version of each task. Response biases to lever press during tests for dealyed responding to S1 were reduced in both symmetrically reinforced tasks, but only in the DD task did such contingencies produce consistently poorer performance in responding to either S1 in Experiment 1. Declines in accuracy of peformance that occurred in both experiments were greater to the visual than to the auditory S1 only in the DMTS tasks with increased intevals between S1 and S2. A third experiment, in which rats had to respond to S2 if it matched S1 (DMTS) or if S2 mismatched S1 (DMmTS), was carried out. Modality of S1 similarly affected accuracy of delayed responding in each task, as in the first two experiments. Methodological and theoretical implications of these results are discussed in terms of Honig and Thompson''s (1982) dual-process theory of working memory.