Douglas S. Grant

Pigeons' memory for event duration: Differences between choice and successive matching tasks☆

Abstract Two experiments examined memory for event duration in pigeons using choice and successive matching tasks. In the choice task, two stimuli were presented following the sample, and the correct choice depended upon the duration of the preceding sample. In the successive task, only one of two stimuli was presented following the sample, and whether responding was or was not reinforced depended upon the duration of the preceding sample. In the first experiment, a successive matching task was employed and accuracy decreased at an equivalent rate as a function of delay on short-sample and long-sample trials, regardless of whether the events which differed in duration were samples of houselight or food. The second experiment, which employed as samples different durations of houselight, replicated this result both in naive subjects and in subjects previously trained in a choice matching task. In addition, it was found that naive subjects which were trained and tested in a choice matching task demonstrated a more rapid decrease in accuracy on long-sample trials than on short-sample trials as a function of delay; a result which has been obtained in several previous studies and has been referred to as the “choose-short” effect. It was also found, however, that the choose-short effect did not occur in the choice matching task if subjects had been trained and tested previously in the successive matching task. It was concluded that pigeons employ different coding strategies in matching to sample duration in the choice and successive tasks.

Symmetrical and asymmetrical coding of food and no-food samples in delayed matching in pigeons

The finding that retention functions decrease more rapidly on trials initiated by a food sample than on trials initiated by a no-food sample was replicated. This result was also obtained for samples that do not contrast hedonically , but that are characterized by the presence versus absence of an event. These results are consistent with an asymmetrical-coding model in which it is held that (a) a no-food sample is not coded, (b) responding to the comparison associated with a no-food sample occurs as a default, and (c) the default response is overridden by the presence in working memory of a code activated by the presentation of food

Analogical and nonanalogical coding of samples differing in duration in a choice-matching task in pigeons.

When trained in a symbolic choice-matching task involving short (2-s) and long (10-s) durations as samples, pigeons tend to choose the comparison associated with a short sample as delay increases (choose-short effect). The present experiments showed that the choose-short effect can be eliminated by training in many-to-one (MTO) procedures in which 2 or more sets of sample stimuli are associated with 1 set of comparison stimuli. It is concluded that (a) the choose-short effect results from a process of subjective shortening that occurs during a delay if duration samples are coded analogically and (b) samples of duration are coded nonanalogically in at least some MTO mapping arrangements

Stimulus Control of Information Processing in Pigeon Short-Term Memory.

Abstract Six pigeons were trained initially on a delayed successive matching-to-sample task using red and green fields as sample and test stimuli. Following acquisition, each sample was followed either by a vertical line (“remember” cue), which indicated that sample memory would be tested, or by a horizontal line (“forget” cue), which indicated that sample memory would not be tested. During the experiments, sample memory on forget trials was tested occasionally. A series of five experiments revealed: (a) better retention on remember trials than on forget trials, (b) increased effectiveness of a forget cue when it followed closely sample offset, (c) more rapid forgetting over a retention interval ranging from 3 to 6 sec on forget trials than on remember trials, (d) a “cancellation” effect in which a remember cue which followed immediately the offset of a forget cue attenuated markedly the effectiveness of the forget cue, and (e) an “insulation” effect in which the effectiveness of a forget cue was reduced considerably when presented after a remember cue. It was concluded that pigeons actively process or rehearse the sample memory during the retention interval.

A differential outcomes effect using biologically neutral outcomes in delayed matching-to-sample with pigeons

The differential outcomes effect (DOE) pertains to enhanced conditional discrimination performance if each correct stimulus-choice sequence is always followed by a different outcome (e.g., food vs. water) compared to when each correct sequence is followed equally often by either outcome. The present experiments sought evidence of a DOE in pigeons, using biologically neutral outcomes. Experiment 1 replicated findings with rats demonstrating that a DOE can occur when one outcome is a biologically neutral light and the other is the absence of that light. Experiment 2 extended these findings by demonstrating a DOE when two biologically neutral outcomes of similar sensory and associative properties were employed.

Mediated transfer testing provides evidence for common coding of duration and line samples in many-to-one matching in pigeons

A three-phase transfer design was used to determine whether pigeons use a single, common code to represent line and duration samples that are associated with the same comparison stimulus. In Phase 1, two sets of samples (two lines and two durations) were associated with either a single set of comparisons (Group MTO, many-to-one) or with different sets of comparisons (Group OTO, one-to-one). In Phase 2, one set of samples was associated with a new set of comparisons. In Phase 3 (transfer test), the alternate set of samples was substituted for the Phase 2 samples. Group MTO, but not Group OTO, demonstrated immediate transfer. It was concluded that associating a line and a duration sample with the same comparison stimulus results in representation of those samples by a single code.

Processing of empty and filled time intervals in pigeons

Pigeons were trained initially with 2- and 8-sec empty or filled intervals as sample stimuli. Interval onset and termination was signaled by 1-sec start and stop markers. Following retention and psychophysical testing, both groups were trained with the alternative type of interval, and the tests were repeated. Group empty-first demonstrated a choose-long effect with both empty and filled intervals. Group filled-first demonstrated a weak (and nonsignificant) choose-short effect with filled intervals and a robust choose-long effect with empty intervals. Both groups tended to time the markers and to add that duration to the sample duration only on filled-interval trials. Initial training with empty intervals alters the way pigeons process temporal information on filled-interval trials, whereas initial training with filled intervals has little effect on the processing of temporal information on empty-interval trials.

Delayed alternation in the rat: Effect of contextual stimuli on proactive interference☆

Abstract The effect of contextual stimuli on proactive interference in the rat was investigated in three experiments. The task was a modified delayed alternation procedure in which the rat was initially forced to one side of a T maze (the interfering forcing) followed by a forcing to the opposite side (the target forcing). A free choice run followed the target forcing in which a turn in either direction could be made. In order to obtain reinforcement on the free choice run, the rat was required to turn in the direction opposite that of the target forcing (i.e., in the same direction as the interfering forcing). Manipulation of the context prevailing during the interfering forcing, target forcing, and free choice runs revealed that free choice accuracy was (a) greater when the interfering and target forcings were conducted in different contexts than when they were conducted in the same context and (b) influenced only slightly, if at all, by whether the free choice context coincided with that prevailing during the interfering forcing or with that prevailing during the target forcing.

Sources of visual interference in delayed matching-to-sample with pigeons.

In two experiments, independent groups of pigeons were trained on an identity matching task involving line orientations as sample and comparison stimuli. For some birds an overhead houselight was illuminated continuously throughout each training session. For other birds the houselight was never illuminated during training sessions. During subsequent testing, the lighting conditions during the delay were the same as in training on some trials, but on other trials they were opposite those of training during either the entire delay (Experiment 1) or during a portion of the delay (Experiment 2). In birds trained with the houselight off, turning the houselight on during the delay produced a large and enduring disruption in matching accuracy. On the other hand, in birds trained with the houselight on, turning the houselight off during the delay produced only a moderate and temporary disruption in matching accuracy. These findings are inconsistent with the prevailing view that retroactive interference in pigeons is a function of a change in illumination level relative to that which prevailed during training. In pigeons, as in monkeys, sustained retoactive interference effects obtain only when the level of illumination is increase during the delay interval.

Samples of stimuli, responses, and reinforcers: Effect of incongruent sample type, serial position, and mode of presentation

In two matching-to-sample experiments, pigeons’ performance with samples of stimuli (red and green), number of responses (1 and 20), and reinforcers (food and no food) was assessed. Samples of red, 20 responses, and food were associated with the red comparison stimulus, and samples of green, 1 response, and no food were associated with the green comparison stimulus. On interference trials, three sample types were presented on each trial, and two of the samples (congruent) were associated with the correct comparison and the third sample (incongruent), with the incorrect comparison. Performance on interference trials was compared with that on control trials in which either two (Experiment 1) or three (Experiment 2) congruent samples were presented. It was found that presentation of an incongruent sample reduced matching accuracy markedly, and about equally, whether samples were presented successively or in compound. Although the type of sample that was incongruent was without effect, matching accuracy declined strongly as the recency of the incongruent sample increased. Serial position of the incongruent sample also influenced the shape of the retention function on interference trials. Presentation of the incongruent sample either first or second resulted in accuracy decreasing across the retention interval, whereas presentation of the incongruent sample last in the input sequence resulted in increasing accuracy across the retention interval. The theoretical implications of the findings are considered.