Errors in pigeons’ memory for number of events

Abstract

A number of studies have employed a delayed matchingto-sample procedure to investigate memory for event duration and memory for number of events in pigeons. In a typical experiment on memory for event duration, trials begin with the presentation of a continuous signal (e.g., the overhead houselight) for either a short (e.g., 2 sec) or a long (e.g., 8 sec) duration. Termination of the sample is followed immediately by the presentation of two comparison stimuli. A peck to one comparison stimulus (e.g., red) is reinforced if the sample duration was short, whereas a peck to the other comparison stimulus (e.g., green) is reinforced if the sample duration was long. Following acquisition, memory for time is assessed by inserting delay intervals of varying length between the termination of the sample and the onset of the comparisons. Previous research has shown that at delays longer than baseline training, pigeons continue to respond with high accuracy following the short-duration sample, whereas accuracy of responding to the long-duration sample drops to well below 50% correct (Fetterman, 1995;Gaitan & Wixted, 2000;Grant, 1993;Kelly & Spetch, 2000;Kraemer, Mazmanian, & Roberts, 1985;Santi, Bridson, & Ducharme, 1993;Santi, Ducharme, & Bridson, 1992;Sherburne, Zentall, & Kaiser, 1998;Spetch, 1987;Spetch & Rusak, 1989Spetch & Wilkie, 1983). This result is commonly known as the choose-short effect, because pigeons show a bias to peck the comparison stimulus that is correct for the shortduration sample as the delay interval is extended beyond that used in training.Similar experiments have been conducted to assess memory for number of events in pigeons. In this case, the sample stimulus consists of a small number or a large number of equally spaced light flashes, each lasting 200 msec. For example, the small sample would consist of two flashes of light in 4 sec (2f/4sec), whereas the large sample would consist of eight flashes of light in 4 sec (8f/4sec). A peck to one comparison stimulus (e.g., red) is reinforced if the sample number was small, whereas a peck to the other comparison stimulus (e.g., green) is reinforced if the sample number was large. Roberts, Macuda, and Brodbeck (1995) used this type of procedure and found that, during delay testing, pigeons continued to respond with high accuracy following the samples consisting of a small number of light flashes (2f/4sec), whereas the accuracy of responding following the large number (8f/4sec) dropped to well below 50% correct. Fetterman (2000) reported a similar finding with rates of stimulus change (slow vs. fast) serving as sample stimuli. Delay tests indicated a bias to respond to the comparison alternative associated with the slow sample. Fetterman (2000) hypothesized that rate discriminations were based on the number of sample events (light flashes) and that the choose-slow bias represented an instance of a choosesmall effect. Thus, memory for number reveals a choosesmall effect, just as memory for time reveals a chooseshort effect. This research was supported by...

Pigeons’ memory for sequences of light flashes when gap duration is an unreliable discriminative cue

“…Spetch and Wilkie referred to this result as the choose-short effect, because pigeons show a bias to peck the comparison stimulus associated with the short sample as the delay interval is extended beyond baseline training. This result has been reported in a large number of studies (Fetterman, 1995;Gaitan & Wixted, 2000;Grant, 1993Grant, , 2006Grant & Kelly, 1996Grant & Spetch, 1991, 1993Grant, Spetch, & Kelly, 1997;Kelly & Spetch, 2000;Kraemer, Mazmanian, & Roberts, 1985;Santi, Bridson, & Ducharme, 1993;Santi, Ducharme, & Bridson, 1992;Sherburne, Zentall, & Kaiser, 1998;Spetch, 1987;Spetch & Rusak, 1989.Memory for number of events has also been studied by training pigeons in a DSMTS procedure (Fetterman, 2000;Fetterman & MacEwen, 1989;Hope & Santi, 2004;Roberts, Macuda, & Brodbeck, 1995;Santi & Hope, 2001). In the studies by Roberts et al and Santi and Hope, pigeons were presented with sample stimuli consisting of two flashes of light in 4 sec (2f/4) or eight flashes of light in 4 sec (8f/4).…”

“…Pigeons' memory for number was assessed by inserting delay intervals of varying length between termination of the sample sequence and the onset of the comparison stimuli. At delays greater than the baseline training delay, pigeons responded with higher accuracy following the few sample, whereas response accuracy following the many sample dropped well below chance (Roberts et al, 1995;Santi & Hope, 2001). This result is referred to as the choose-few effect because pigeons are biased to respond to the comparison stimulus that was correct for the smaller number sample as the delay interval is extended beyond baseline training.…”

“…Thus, pigeons may have coded the number samples by timing and summing f lash duration, as opposed to using an event switch to count the number of feeder light flashes. If this interpretation is correct, then the studies by Roberts et al (1995), Santi and Hope (2001), and Hope and Santi (2004) were not studies examining memory for number, but studies of memory for event duration. Table 1 identifies the temporal and numerical features of the few and many samples used in previous studies of memory for number in the pigeon.…”

“…In accord with this account, both the choose-few effect and the choose-short effect have been shown to depend on the baseline training delay. If a constant delay interval (e.g., 5 or 10 sec) is used during baseline training in either a number (Santi & Hope, 2001) or a duration (Spetch, 1987) discrimination, then the introduction of shorter delay intervals during testing results in a choosemany effect for number samples and a choose-long effect for duration samples. Hope and Santi (2004) examined pigeons' memory for number of feeder light flashes, varying the ambient chamber illumination during the intertrial interval (ITI) and the delay interval.…”

“…Memory for number of events has also been studied by training pigeons in a DSMTS procedure (Fetterman, 2000;Fetterman & MacEwen, 1989;Hope & Santi, 2004;Roberts, Macuda, & Brodbeck, 1995;Santi & Hope, 2001). In the studies by Roberts et al and Santi and Hope, pigeons were presented with sample stimuli consisting of two flashes of light in 4 sec (2f/4) or eight flashes of light in 4 sec (8f/4).…”